Ferroelectric Devices 2nd Edition

Ferroelectric
Devices
Second Edition
© 2010 by Taylor & Francis Group, LLC

Ferroelectric
Devices
Second Edition
Kenji Uchino
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works

Version Date: 20110720
International Standard Book Number-13: 978-1-4398-0376-9 (eBook - PDF)
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-
ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may
rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or uti-
lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy-
ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the
publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923,
978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com

Contents
Preface...............................................................................................................................................xi
Suggested Teaching Schedule ........................................................................................................ xiii
Prerequisite Knowledge ................................................................................................................... xv
Author ............................................................................................................................................xvii
List of Symbols ...............................................................................................................................xix
Chapter 1 General View of Ferroelectrics ....................................................................................1

1.1 Crystal Structure and Ferroelectricity .................................................................1
1.2 Origin of Spontaneous Polarization ....................................................................4
1.3 Origin of Field-Induced Strain ............................................................................7
1.4 Electro-Optic Effect .......................................................................................... 11
1.5 Example of Ferroelectrics ................................................................................. 17
1.5.1 Barium Titanate ..................................................................................... 17
1.5.2 PZT ........................................................................................................ 18
1.5.3 Polyvinylidene Difluoride (PVDF) ........................................................ 19
1.5.4 Pb-Free Piezoceramics ..........................................................................20
1.6 Applications of Ferroelectrics ...........................................................................20
References .................................................................................................................. 23
Chapter 2 Mathematical Treatment of Ferroelectrics .................................................................25

2.1 Tensor Representation of Physical Properties ...................................................25
2.1.1 Fundamentals of Tensor .........................................................................25
2.1.2 Tensor Representation ............................................................................26
2.1.3 Crystal Symmetry and Tensor Form...................................................... 27
2.1.4 Reduction of the Tensor (Matrix Notation)............................................ 33
2.1.5 Matrix Notation of Electrostrictive Coefficients ...................................34
2.2 Phenomenology of Ferroelectricity ................................................................... 42
2.2.1 Fundamentals of Phenomenology.......................................................... 42
2.2.2 Landau Theory of the Phase Transition ................................................ 43
2.2.3 Phenomenology of Electrostriction ....................................................... 47
2.2.4 Converse Effects of Electrostriction ...................................................... 50
2.2.5 Temperature Dependence of Electrostriction ........................................ 50
2.3 Phenomenology of Antiferroelectricity............................................................. 51
2.3.1 Antiferroelectrics ................................................................................... 51
2.3.2 Phenomenology of Antiferroelectrics .................................................... 52
2.4 Ferroelectric Domain Contributions ................................................................. 54
2.4.1 Ferroelectric Domain Reorientation ...................................................... 54
2.4.2 Uchida–Ikeda Model ............................................................................. 57
2.4.3 Crystal Structure and Coercive Field .................................................... 61
2.4.4 Hysteresis Estimation Program ............................................................. 61
2.4.5 Domain Engineering .............................................................................64
References .................................................................................................................. 73

vi Contents
Chapter 3 Designing with Materials and Devices and Fabrication Processes............................ 75

3.1 Material Designing ............................................................................................ 75
3.1.1 Composition Selection............................................................................ 75
3.1.2 Dopant Effects on Ferroelectricity......................................................... 77
3.1.2.1 Domain Wall Stability ............................................................. 78
3.1.2.2 Crystallographic Deficiencies..................................................80
3.1.3 High-Power Characteristics ................................................................... 82
3.2 Fabrication Processes of Ceramics.................................................................... 85
3.2.1 Preparation of Ceramic Powders ........................................................... 85
3.2.1.1 Solid-State Reaction ................................................................ 85
3.2.1.2 Coprecipitation ........................................................................ 86
3.2.1.3 Alkoxide Hydrolysis ................................................................ 86
3.2.2 Sintering Process ................................................................................... 87
3.2.3 Single Crystal Growth ...........................................................................90
3.2.3.1 Quartz, LN, and LT .................................................................90
3.2.3.2 PZN–PT, PMN–PT, and PZT .................................................90
3.2.4 Templated Grain Growth .......................................................................90
3.3 Device Designing .............................................................................................. 91
3.3.1 Single Disks ........................................................................................... 91
3.3.2 Multilayers .............................................................................................92
3.3.3 Unimorphs/Bimorphs ............................................................................94
3.3.4 Flextension/Hinge-Lever Amplification Mechanisms .......................... 98
3.3.5 Flexible Composites...............................................................................99
3.3.6 Thin/Thick Films...................................................................................99
3.3.6.1 Film Manufacturing Techniques .............................................99
3.3.6.2 MEMS Applications.............................................................. 101
3.3.6.3 Constraints in Thin/Thick Films .......................................... 102
3.4 Size Effect on Ferroelectricity......................................................................... 102
3.4.1 Grain Size Effect on Ferroelectricity ................................................... 103
3.4.2 3-D Particle Size Effect on Ferroelectricity ........................................ 105
References ................................................................................................................ 112
Chapter 4 High-Permittivity Dielectrics ................................................................................... 115

4.1 Ceramic Capacitors ......................................................................................... 115
4.2 Chip Capacitors ............................................................................................... 116
4.2.1 Thin Multilayer Capacitor ................................................................... 116
4.2.2 Base-Metal Internal Electrode ............................................................ 117
4.2.2.1 Barium Titanate–Based Chip................................................ 117
4.2.2.2 PbTiO3-Based Chip ............................................................... 117
4.3 Hybrid Substrates ............................................................................................ 119
4.4 Relaxor Ferroelectrics ..................................................................................... 121
4.4.1 High Permittivity ................................................................................. 121
4.4.2 Diffuse Phase Transition ..................................................................... 122
4.4.3 Dielectric Relaxation ........................................................................... 124
References ................................................................................................................ 130
Chapter 5 Ferroelectric Memory Devices................................................................................. 131

5.1 DRAM ............................................................................................................. 131
5.1.1 Principle of DRAM .............................................................................. 131
5.1.2 Ferroelectric DRAM ............................................................................ 135

Contents vii
5.2 Nonvolatile Ferroelectric Memory .................................................................. 136

5.2.1 FeRAM (Inversion Current Type) ....................................................... 136
5.2.2 MFSFET .............................................................................................. 139
References ................................................................................................................ 142
Chapter 6 Pyroelectric Devices ................................................................................................ 145

6.1 Pyroelectric Materials ..................................................................................... 145
6.1.1 Pyroelectric Effect ............................................................................... 145
6.1.2 Responsivity ......................................................................................... 148
6.1.3 Figures of Merit ................................................................................... 150
6.2 Temperature/Infrared Sensors......................................................................... 152
6.3 Infrared Image Sensors ................................................................................... 152
6.4 Electrocaloric Devices..................................................................................... 154
6.4.1 Electrocaloric Effect ............................................................................ 154
6.4.1.1 Phenomenological Formula ................................................... 154
6.4.1.2 Electrocaloric Materials ........................................................ 157
References ................................................................................................................ 160
Chapter 7 Piezoelectric Devices ............................................................................................... 161

7.1 Piezoelectric Materials and Properties ............................................................ 161
7.1.1 Piezoelectric Figures of Merit .............................................................. 161
7.1.2 Piezoelectric Materials ......................................................................... 167
7.1.2.1 Single Crystals ....................................................................... 168
7.1.2.2 Polycrystalline Materials ....................................................... 168
7.1.2.3 Relaxor Ferroelectrics............................................................ 172
7.1.2.4 Polymers ................................................................................ 172
7.1.2.5 Composites ............................................................................ 173
7.1.2.6 Thin Films ............................................................................. 173
7.2 Pressure Sensors/Accelerometers/Gyroscopes ................................................ 174
7.3 Piezoelectric Vibrators/Ultrasonic Transducers .............................................. 177
7.3.1 Piezoelectric Resonance ....................................................................... 177
7.3.1.1 Piezoelectric Equations .......................................................... 177
7.3.1.2 Electromechanical Coupling Factor ...................................... 177
7.3.1.3 Longitudinal Vibration Mode ................................................ 180
7.3.2 Equivalent Circuits of Piezoelectric Vibrators..................................... 184
7.3.3 Piezoelectric Vibrators ......................................................................... 186
7.3.4 Ultrasonic Transducers ........................................................................ 187
7.3.4.1 Ultrasonic Imaging ................................................................ 187
7.3.4.2 Sonochemistry ....................................................................... 188
7.3.5 Resonators/Filters ................................................................................ 189
7.4 Surface Acoustic Wave Devices ...................................................................... 191
7.5 Micromass Sensors .......................................................................................... 193
7.5.1 Biosensor .............................................................................................. 193
7.5.2 Viscosity Sensor ................................................................................... 194
7.6 Piezoelectric Transformers .............................................................................. 194
7.7 Piezoelectric Actuators .................................................................................... 198
7.7.1 Ceramic Actuator Materials ................................................................. 198
7.7.2 Actuator Designs ..................................................................................200
7.7.3 Drive/Control Techniques .................................................................... 201

viii Contents
7.7.4
Device Applications .............................................................................207
7.7.4.1 Servo Displacement Transducers ...........................................207
7.7.4.2 Pulse-Drive Motors ................................................................208
7.8 Ultrasonic Motors ............................................................................................ 212
7.8.1 Ultrasonic Motors (USM) .................................................................... 212
7.8.2 Classification and Principles of Ultrasonic Motors ............................. 214
7.8.2.1 Standing-Wave Type .............................................................. 214
7.8.2.2 Propagating-Wave Type ......................................................... 215
7.8.3 Standing Wave–Type Motors ............................................................... 216
7.8.3.1 Rotary Motors........................................................................ 216
7.8.3.2 Linear Motors ........................................................................ 218
7.8.4 Propagating Wave–Type Motors .......................................................... 219
7.8.4.1 Linear Motors ........................................................................ 219
7.8.4.2 Rotary Motors ....................................................................... 220
7.8.4.3 Comparison of the Various Ultrasonic Motors ..................... 223
7.8.5 Smooth Impact Drive Mechanism (SIDM) .........................................224
7.9 Piezoelectric Energy Harvesting .....................................................................224
7.9.1 Piezoelectric Passive Damping to Energy Harvesting .........................224
7.9.2 High-Energy Harvesting (~W) ............................................................. 225
7.9.3 Low-Energy Harvesting (~mW) ........................................................... 227
References ................................................................................................................ 232
Chapter 8 Electro-Optic Devices .............................................................................................. 235

8.1 Electro-Optic Effect—Review ........................................................................ 235
8.2 Transparent Electro-Optic Ceramics .............................................................. 235
8.2.1 (Pb,La)(Zr,Ti)O3 ................................................................................... 236
8.2.2 PMN–PT.............................................................................................. 239
8.2.3 Pb(Zn1/3Nb2/3)O3 .................................................................................. 241
8.3 Bulk Electro-Optic Devices ............................................................................ 242
8.3.1 Ferpic ................................................................................................... 242
8.3.2 Eye Protection Application .................................................................. 243
8.3.3 Stereo TV Application ......................................................................... 243
8.3.4 Two-Dimensional Displays ..................................................................244
8.3.4.1 Fabrication Process of the 2-D Display .................................244
8.3.4.2 Characteristics of the Light Valve Array ..............................246
8.3.4.3 Construction of the Image Projector ..................................... 247
8.3.5 KTN Single Crystal Optical Switch .................................................... 251
8.3.6 Laser Beam Scanner ............................................................................ 252
8.4 Waveguide Modulators .................................................................................... 252
8.4.1 LiNbO3 Waveguide .............................................................................. 252
8.4.2 PZT Thin Film Waveguide .................................................................. 254
References ................................................................................................................ 259
Chapter 9 PTC Materials .......................................................................................................... 261

9.1 Mechanism of PTC Phenomenon .................................................................... 261
9.1.1 PTC Phenomenon ................................................................................. 261

Contents ix
9.1.2 Mechanism of the PTC Phenomenon................................................... 261

9.1.2.1 Semiconducting Properties .................................................... 261
9.1.2.2 Schottky Barrier Effect ......................................................... 263
9.1.2.3 Experimental Verification ..................................................... 265
9.2 PTC Thermistors .............................................................................................266
9.3 Grain Boundary Layer Capacitors................................................................... 269
References ................................................................................................................ 272
Chapter 10 Composite Materials ................................................................................................ 273

10.1 Connectivity................................................................................................... 273
10.2 Composite Effects.......................................................................................... 274
10.2.1 Sum Effects ...................................................................................... 274
12.2.2 Combination Effects ........................................................................ 275
10.2.3 Product Effects ................................................................................. 276
10.3 PZT:Polymer Composites .............................................................................. 277
10.3.1 Piezoelectric Composite Materials ................................................... 277
10.3.2 Principle of PZT:Polymer Composites ............................................. 278
10.3.3 Theoretical Models for 0-3 Composites ........................................... 281
10.3.4 Advanced PZT:Polymer Composites................................................ 283
10.4 Composite Dampers and Energy Harvesters .................................................284
10.4.1 Piezoelectric Composite Dampers....................................................284
10.4.2 Piezoelectric Composite Energy Harvesting.................................... 288
10.5 Magnetoelectric Sensors................................................................................ 290
References ................................................................................................................ 296
Chapter 11 Future of Ferroelectric Devices................................................................................ 297

11.1 Market Share .................................................................................................. 297
11.1.1 Market Share of Ferroelectric Devices ............................................. 297
11.1.2 Market Share of Piezoelectric Actuators .......................................... 298
11.2 Reliability Issues ............................................................................................300
11.2.1 Materials Improvements ...................................................................300
11.2.2 Reliability of Devices ....................................................................... 301
11.2.3 Drive/Control Techniques ................................................................ 303
11.2.4 Safety Systems.................................................................................. 303
11.3 Development of Bestselling Devices .............................................................304
11.3.1 Three Creativities .............................................................................304
11.3.2 Technological Creativity...................................................................304
11.3.2.1 Discovery of a New Function or Material ........................304
11.3.2.2 Performance Improvement ...............................................308
11.3.3 Product Planning Creativity ............................................................. 314
11.3.3.1 Seeds and Needs ............................................................... 314
11.3.3.2 Innovation Obstacle in Technology Management ............ 316
11.3.3.3 Development Pace ............................................................. 316
11.3.3.4 Specifications .................................................................... 316
11.3.3.5 Product Design Philosophy............................................... 318
11.3.3.6 Smart Systems................................................................... 318

x Contents
11.3.4 Marketing Creativity ........................................................................ 319

11.3.4.1 Choose Your Customers.................................................... 320
11.3.4.2 Narrow Your Focus ........................................................... 322
11.3.4.3 Dominate Your Market ..................................................... 326
11.3.5 Business Strategy.............................................................................. 327
11.3.5.1 Analysis Tools ................................................................... 328
11.3.5.2 Internal Environmental Analysis ...................................... 329
11.3.5.3 External Environmental Analysis ..................................... 331
References ................................................................................................................ 337
Index .............................................................................................................................................. 339

Preface
Ferroelectrics can be used in various devices such as high-permittivity dielectrics, pyroelectric sen-
sors, piezoelectric devices, electro-optic devices, and positive temperature coefficient (PTC) com-
ponents. The concerned industries are constantly producing large amounts of simple devices such
as ceramic capacitors, piezoelectric igniters, buzzers, and PTC thermistors. However, ferroelectric
devices have, until now, failed to reach commercialization in more functional cases. In light sen-
sors, for example, semiconductive materials are superior to ferroelectrics in terms of response speed
and sensitivity. Magnetic devices and semiconductive “Flash” memories are much more popular in
the memory field, and liquid crystals are typically used for optical displays. Ferroelectric devices
often fail to be developed when other competitive materials exist. However, taking into consider-
ation recent strides that have been made at commercialization, I have not given up on ferroelectric
devices. I believe that these earlier failures were mainly due to a lack of systematic accumulation of
fundamental knowledge of the materials and developmental experiences of the devices.
During my 35-year teaching period on ferroelectric devices, I was unable to find any book dedi-
cated exclusively to this field, other than some professional publications such as multiauthored paper
collections. Hence, I decided to collect my lecture notes, including my device-development phi-
losophy, and write this comprehensive book. This book provides an introduction to the theoretical
background of ferroelectric devices, practical materials, device designs, drive/control techniques,
and typical applications, and looks forward to future progress in this field. Although ferroelectricity
was discovered a long time ago, device development is relatively new and interdisciplinary, and it
is therefore almost impossible to cover all the recent studies in one single book. Therefore, in this
second edition, I have focused on the important ideas to understand how to design and develop
ferroelectric devices by adding some up-to-date device developments to the original devices.
This book is organized as follows. Chapter 1 provides a general introduction while Chapter 2
provides more of a theoretical background. Chapter 3 deals with the practical designing and manu-
facturing of the devices, including recent single-crystal and thin-film processes. Chapter 4 elaborates
on capacitor applications while Chapters 5 and 6 deal with thin- and thick-film applications, i.e.,
ferroelectric memory devices and pyroelectric devices, respectively. A new section on electrocaloric
devices has been inserted in the last part of Chapter 6. Chapter 7 deals with piezoelectric actua-
tors and ultrasonic motors as well as acoustic transducers and piezoelectric sensors. Piezoelectric
energy harvesting is a new piezoelectric topic. Chapter 8 provides a description of optical devices
such as light valves, displays, wave guides, and bulk photovoltaic devices. In Chapters 9 and 10, the
basic concepts of PTC materials and composite materials are explained and their device applica-
tions are demonstrated. Finally, Chapter 11 discusses the future of ferroelectric devices, that is, the
market size and the remaining reliability/lifetime issues, and the author’s strategy for developing
best-selling devices.
One decade has passed since the first edition appeared. During this period, the book was for-
tunately a commercial success and has been well received as a standard in this field. Although the
fundamental principles, including Maxwell equations, phenomenology, and piezoelectric relations,
have not been altered, the applications have been updated in keeping with the times.
The second edition aims primarily at including up-to-date research and developmental trends,
reflecting nano- and optical technologies, environmental regulation, and alternative energy
sources. The new sections in this edition include (1) Pb-free piezoelectrics, (2) size effects on fer-
roelectricity, (3) electrocaloric devices, (4) micro-mass sensors, (5) piezoelectric energy harvesting,
(6) light valves and scanners, and (7) multiferroic devices, which, in turn, include magnetoelectric
sensors.
xi

xii Preface
Even for traditional devices, such as capacitors, piezoelectric speakers, clocks, and transformers,
old data have been updated with an eye to the future.
This edition is a more complete book; in addition to the chapter summaries, example problems,
and chapter problems,
• FAQs (frequently asked questions)—to answer common in-class questions,

• Lab Demonstrations—to provide effective in-class demos for practical understanding
• Check Points—to have a fundamental understanding check for students
have been newly inserted in every chapter. Over the years, all of these had accumulated in my
notebook, courtesy of students’ feedback and comments. These three additions will help the
reader correctly understand the essence of this book. Furthermore, due to strong requests by junior
instructors, model solutions have been added for the check point and chapter problems in a separate
solutions manual. I usually use check points for quizzes, example problems for homework, and
chapter problems for mid-term and final exams. I hope this will make it easier for the instructor to
prepare exam and homework problems.
Most errors in mathematical formulas, English expressions, and typos have been completely
revised and fixed in this edition.
This book was written for graduate students, undergraduate junior and senior levels, and industry
engineers studying or working in the fields of electronic materials, optical materials and communi-
cations, precision machinery, and robotics. Although this book is designed as a course with thirty
75-min lectures, readers can learn by themselves aided by the examples and problems.
For those who seek more detailed information on smart piezoelectric actuators and sensors,
Micromechatronics by K. Uchino and the late Jayne R. Giniewicz, published by CRC/Dekker in
2003, is recommended. Further, FEM and Micromechatronics with ATILA Software, published by
CRC Press in 2008, is a perfect manual for practical device designing.
Even though I am the sole author of this book, it includes the contributions of many others.
I express my gratitude to my ICAT center faculty members who have generously given me their
advice and help during the writing, particularly Professor Uma Belegundu at Lock Haven University,
who worked out all the problems. Specific acknowledgment and condolences are also dedicated to
my former student at Indiana University of Pennsylvania, Professor Jayne Giniewicz, who reviewed
and criticized the entire manuscript and made linguistic corrections, but passed away unexpectedly
at a very young age. Finally, my wife Michiko’s constant assistance is greatly appreciated with my
deepest love.
Critical review and corrections to this book will be highly appreciated. Send your feedback
directly to Kenji Uchino at 134 Materials Research Laboratory, The Pennsylvania State University,
University Park, PA 16802-4800; fax: 814-863-6734; e-mail: KenjiUchino@PSU.EDU.
Kenji Uchino

Suggested Teaching Schedule
(75 min × 30 times per semester)
0. Course Explanation and Prerequisite Knowledge Check 1 Time

1. General View of Ferroelectrics 4 Times
2. Mathematical Treatment of Ferroelectrics 4 Times
3. Designing with Materials and Devices and
Fabrication Processes 2 Times
4. High-Permittivity Dielectrics 2 Times
5. Ferroelectric Memory Devices 1 Time
6. Pyroelectric Devices 2 Times
7. Piezoelectric Devices 5 Times
8. Electro-Optic Devices 2 Times
9. PTC Materials 1 Time
10. Composite Materials 2 Times
11. Future of Ferroelectric Devices 1 Time
Lab Demonstration 2 Times

Review/Q&A 1 Time
xiii

Prerequisite Knowledge
In order to understand ferroelectric devices, some prerequisite knowledge is expected. Try to solve
the following questions without seeing the answers on the next page.
Q1 Describe the definitions of elastic stiffness c and compliance s, using a stress X–strain
x relation.
Q2 Indicate a shear stress X4 on the following square.
Q3 Describe the sound velocity v in a material with mass density ρ and elastic compliance sE.
Q4 Calculate the capacitance C of a capacitor with area S and electrode gap t filled with a material
of relative permittivity ε.
Q5 Calculate the polarization P of a material with dipole density N (m−3) of dipole moment
qu (C m).
Q6 Describe the Curie–Weiss law of relative permittivity ε, using a Curie–Weiss temperature T0
and a Curie–Weiss constant C.
Q7 Describe the light velocity in a material with a refractive index n (c: light velocity in
vacuum).
Q8 Indicate the work function in the following energy band of a metal:
Vacuum level
Energy
Fermi level
Inside
of metal Outside
Q9 There is a voltage supply with an internal impedance Z 0. Indicate the external impedance Z1
to obtain the maximum output power.
Q10 Calculate the induced polarization P under an external stress X in a piezoelectric with a piezo-
electric constant d.
xv

xvi Prerequisite Knowledge
ANSWERS
(Correct rate more than 60% of the full score is expected)
Q1 X = cx, x = sX
Q2 X4 = 2X23 = 2φ (rad)
X4
φ
Q3 v = 1 / ρS E [0.5 point for v = 1/ρsE]

Q4 C = ε0ε(S/t) [0.5 point for C = ε(S/t)]
Q5 P = Nqu
Q6 ε = C/(T – T0) [0.5 point for ε = C/T]
Q7 c′ = c/n
Q8 (Work function)
Vacuum level
Work
function
Energy
Fermi level
Inside
Outside
of metal
Q9 Z1 = Z 0
On Z1, the current and voltage are given as V/(Z 0 + Z1) and [Z1/(Z 0 + Z1)]V, leading to the
power:
Power = V 2 Z1/(Z 0 + Z1)2 = V 2/(Z 0 / Z1 + Z1 )2 ≤ (1/4)V 2/Z 0
The maximum is obtained when Z0 / Z1 = Z1 . Hence, Z1 = Z 0.
Z0
V Z1
Q10 P = dX (refer to x = dE)

Author
Kenji Uchino, one of the pioneers in piezoelectric actuators and electro-optic displays, is the director
of the International Center for Actuators and Transducers (ICAT) and professor of electrical engineer-
ing at The Pennsylvania State University. He is currently teaching “ferroelectric devices,” making
use of this book, as well as two other books, Micromechatronics and FEM and Micromechatronics
with ATILA Software, that have been published by CRC Press. He is also the senior vice president
and CTO of Micromechatronics Inc., a spin-off company of ICAT, where he is trying to commer-
cialize ICAT-invented piezoactuators and transducers. He has also started teaching “how to start-up
a high-tech company” for young engineers using a textbook Entrepreneurship for Engineers pub-
lished by CRC Press.
After receiving his PhD from the Tokyo Institute of Technology, Japan, Uchino became a
research associate in the physical electronics department at the same university. He then joined
Sophia University, Japan, as an associate professor of physics in 1985. Subsequently, in 1991, he
moved to Pennsylvania. He was also involved with the Space Shuttle Utilizing Committee at the
National Space Development Agency (NASDA), Japan, during 1986–1988, and was the vice presi-
dent of NF Electronic Instruments, United States, during 1992–1994. Uchino has an additional mas-
ter’s degree in business administration from St. Francis University, Loretto, Pennsylvania. He has
served as a consultant for more than 100 Japanese, U.S., and European industries to commercialize
piezoelectric actuators and electro-optic devices. He is the chairman of the Smart Actuator/Sensor
Study Committee that is partly sponsored by the Japanese government, Ministry of International
Trade and Industry (MITI). He is also the associate editor of the Journal of Materials Technology
(Matrice Technology) and an editorial board member of the Journal of Ferroelectrics (Gordon &
Breach) and the Journal of Electroceramics (Kluwer Academic). He also served as an administra-
tive committee member for the IEEE Ultrasonics, Ferroelectrics, and Frequency Control Society
during 1998–2000.
Uchino’s research interests are in solid-state physics—especially dielectrics, ferroelectrics, and
piezoelectrics, including basic research on materials, device designing, and fabrication processes;
he is also interested in the development and commercialization of solid-state actuators and dis-
plays for precision positioners, ultrasonic motors, projection-type TVs, etc. He has authored 550
papers, 54 books, and 26 patents on the subject of piezoelectric actuators and optical devices. He
has been a fellow of the American Ceramic Society since 1997, and is also a recipient of the SPIE
Smart Product Implementation Award (2007), the R&D 100 Award (2007), the ASME Adaptive
Structures Prize (2005), the Outstanding Research Award from the Penn State Engineering Society
(1996), the Academic Scholarship from Nissan Motors Scientific Foundation (1990), the Best Movie
Memorial Award at the Japan Scientific Movie Festival (1989), and the Best Paper Award from the
Japanese Society of Oil/Air Pressure Control (1987).
Apart from his academic pursuits, Uchino is an honorary member of KERAMOS (National
Professional Ceramic Engineering Fraternity) and has received the Best Movie Memorial Award as
the director/producer of several educational videos on “dynamical optical observation of ferroelectric
domains” and “ceramic actuators” at the Japan Scientific Movie Festival (1989).
xvii

List of Symbols
c Elastic stiffness
C Curie–Weiss constant
D Electric displacement
d, g Piezoelectric coefficients
E Electric field
g Secondary electro-optic coefficient
G1 Gibbs elastic energy
k Electromechanical coupling factor
M, Q Electrostrictive coefficients
n Refractive index
p Pyroelectric coefficient
P Dielectric polarization
Ps Spontaneous polarization
r Primary electro-optic coefficient
s Elastic compliance
T0 Curie–Weiss temperature
TC Curie temperature (phase transition temperature)
v Sound velocity
x Strain
X Stress
xs Spontaneous strain
α Ionic polarizability
γ Lorentz factor
Γ Phase retardation
ε Relative permittivity
ε0 Vacuum permittivity
η Energy transmission coefficient
μ Dipole moment
xix

1 General View of Ferroelectrics
Let us start with the “smartness” of a material. Table 1.1 lists the various effects relating the input
(electric field, magnetic field, stress, heat, and light) with the output (charge/current, magnetiza-
tion, strain, temperature, and light). Conducting and elastic materials, which generate current and
strain outputs, respectively, for the input voltage or stress (well-known phenomena), are sometimes
called “trivial” materials. On the other hand, pyroelectric and piezoelectric materials, which gen-
erate an electric field with the input of heat and stress (unexpected phenomena), respectively, are
called “smart” materials. These off-diagonal couplings have corresponding converse effects, the
electrocaloric and converse piezoelectric effects, and both “sensing” and “actuating” functions can
be realized in the same materials. “Intelligent” materials must possess a “drive/control” or “pro-
cessing” function which is adaptive to changes in environmental conditions, in addition to actuator
and sensing functions. Ferroelectric materials exhibit most of these effects with the exception of the
magnetic phenomena. Thus, ferroelectrics are said to be very “smart” materials.
Ferroelectrics are utilized in various devices such as high-permittivity dielectrics, pyroelectric sen-
sors, piezoelectric devices, electro-optic devices, and positive temperature coefficient (PTC) of resistiv-
ity components. However, ferroelectric devices often fail to be commercialized in areas of application
where competitive materials exist. Light sensors, for example, typically are manufactured from semi-
conductive materials which are superior to ferroelectrics in response speed and sensitivity. Magnetic
devices are much more popular for memory applications, and liquid crystals are typically used for opti-
cal displays. One reason for this is due to the lack of systematic and comprehensive compilation of
knowledge on ferroelectric materials. In this chapter, we will learn the fundamentals of ferroelectricity.
1.1 CRYSTAL STRUCTURE AND FERROELECTRICITY

In the so-called dielectric materials, the constituent atoms are considered to be ionized to a certain
degree and are either positively or negatively charged. In such ionic crystals, when an electric field
is applied, cations are attracted to the cathode and anions to the anode due to electrostatic inter-
action. The electron clouds also deform, causing electric dipoles. This phenomenon is known as
electric polarization of the dielectric, and the polarization is expressed quantitatively as the sum of
the electric dipoles per unit volume (C/m2). Figure 1.1 shows schematically the origin of the electric
polarization. There are three primary contributions: electronic, ionic, and dipole reorientation-
related. The degree to which each mechanism contributes to the overall polarization of the material
depends on the frequency of the applied field. Electronic polarization can follow alternating fields
with frequencies from terahertz (THz) to petahertz (PHz) [(1012−1015 cycle/s), higher than visible
light wave] and ionic polarization responds to frequencies from gigahertz (GHz) to THz [(109–1012
cycle/s), microwave region]. Thus, you should understand that a famous relation between the relative
permittivity ε and the refractive index n:
ε = n2 (1.1)
is valid only when the applied electric field has a frequency on the order of THz or higher. Permanent
dipole reorientation can follow only frequencies from megahertz (MHz) to GHz (106 –109 cycle/s).
This is why ferroelectric materials with permanent dipoles cannot be used for microwave dielectric
materials; their permittivity is typically high at low frequencies (kHz), but decreases significantly
with increasing applied electric field frequency. Frequency dependence of the total polarizability (or
permittivity) is depicted in Figure 1.2.

2 Ferroelectric Devices
TABLE 1.1
Various Effects in Materials
Input Material Output

device
Output Charge Magnet-

Input current ization Strain Temperature Light
Permittivity Elect.-mag. Converse Elec. caloric Elec.-optic

Elec. field
Conductivity effect piezoeffect effect effect
Mag.-elect. Magneto- Mag. caloric Mag. optic

Mag. field Permeability
effect striction effect effect
Piezoelectric Piezomag. Elastic Photoelastic

Stress ___
effect effect constant effect
Pyroelectric Thermal Specific

Heat ___
effect expansion heat
Photovoltaic Refractive
Light Photostriction ___
effect index
Diagonal coupling Sensor

Off-diagonal coupling = Smart material Actuator
Compared with air-filled capacitors, dielectric capacitors can store more electric charge due to
the dielectric polarization P as shown in Figure 1.3. The physical quantity corresponding to the
stored electric charge per unit area is called the electric displacement D, and is related to the electric
field E by the following expression:
D = ε0 E + P = εε 0 E. (1.2)
Here,
ε0 is the vacuum permittivity (=8.854 × 10 −12 F/m)
ε is the material’s relative permittivity (also simply called permittivity or dielectric constant, and
in general is a tensor property)
E=0 E
Electronic
+ – +
polarization –
Ionic
polarization + – + + – +
– +
– + – + – + – + –+ – +
Dipole + –
reorientation – +
+ – – + – + – + – + – +
+ –
FIGURE 1.1 Microscopic origins of the electric polarization.

General View of Ferroelectrics 3
Microwave Infrared Ultraviolet
Total polarizability (real part)

αdipolar
αionic
αelectronic
108 1012 1016

Frequency (Hz)
FIGURE 1.2 Frequency dependence of the polarizability (or permittivity).
–
–σf – – – – – –
+ + + + +σb
+
– – – – –
Dielectric
E + + + + +
– – – – –
–σb
+σf + + + + + +
σt = σb + σf
+
σf : Free charge σb: Bound charge σt: True charge
FIGURE 1.3 Charge accumulation in a dielectric capacitor.
Depending on the crystal structure, the centers of the positive and negative charges may not
coincide even without the application of an external electric field. Such crystals are said to possess
a spontaneous polarization. When the spontaneous polarization of the dielectric is reversed by an
electric field, it is called ferroelectric.
Not every dielectric is a ferroelectric. Crystals can be classified into 32 point groups according to
their crystallographic symmetry, and these point groups can be divided into two classes, one with a
center of symmetry and the other without as indicated in Table 1.2. There are 21 point groups which
do not have a center of symmetry. In crystals belonging to 20 of these point groups [point group
(432) being the sole exception], positive and negative charges are generated on the crystal surfaces
when appropriate stresses are applied. These materials are known as piezoelectrics.
Pyroelectricity is the phenomenon in which, due to the temperature dependence of the sponta-
neous polarization, a change in temperature of the crystal produces electric charges on the surface
of the crystal corresponding to the change of spontaneous polarization. Among the pyroelectric
crystals, those whose spontaneous polarization can be reversed by an electric field (not exceeding
the breakdown limit of the crystal) are called ferroelectrics. There is some experimental ambigu-
ity in this definition. In establishing ferroelectricity, it is necessary to apply an electric field to a
pyroelectric material and experimentally ascertain the polarization reversal.

TABLE 1.2
Crystallographic Classification according to Crystal Centrosymmetry
and Polarity
Crystal System
Rhombo- Ortho- Mono-

Polarity Symmetry Cubic Hexagonal Tetragonal hedral rhombic clinic Triclinic
Centro
(11) m3m m3 6/mmm 6/m 4/mmm 4/m 3m 3 mmm 2/m 1
Nonpolar
(22) 432 622 422
23 6 4 32 222
Noncentro 43m 6m2 42m
(21)
Polar 2
(pyroelectric) 6mm 6 4mm 4 3m 3 mm2 1
(10) m
Note: Inside the bold line are piezoelectrics.
1.2 ORIGIN OF SPONTANEOUS POLARIZATION

Why is it that crystals which, from a consideration of elastic energy, should be stable by being non-
polar, still experience the shifting of cations and anions and become spontaneously polarized? The
reason is briefly explained below. For simplicity, let us assume that dipole moments result from the
displacement of one type of ion A (electric charge q) relative to the crystal lattice. Consider the case
in which the polarization is caused by all the A ions being displaced equally in a lattice.
This kind of ionic displacement can be expected through lattice vibrations at a finite tempera-
ture. Figure 1.4 shows some of the possible eigen lattice vibrations in a perovskite-like crystal.
Figure 1.4a shows an initial cubic (symmetrical) structure, Figure 1.4b is a symmetrically elon-
gated one, Figure 1.4c has coherently shifted center cations, and Figure 1.4d exhibits an anti-
polarized shift of the center cations. If one particular lattice vibration lowers the crystal energy,
the ions will shift and stabilize the crystal structure so as to minimize the energy. Starting from
the original cubic structure (Figure 1.4a), if Figure 1.4b is stabilized, only oxygen octahedra
(a) (b) (c) (d)
FIGURE 1.4 Some possible eigen lattice vibration modes in a perovskite crystal. (a) Initial symmetric cubic,
(b) symmetrically elongated, (c) coherently shifted centre cations, and (d) antipolarized shift of the centre
cations.

+ + + Equipotential
– – –
E-line
r1
+ pi
ri –
E0
+ + +
– – – Dielectric material
FIGURE 1.5 Concept of the local field. Eloc is given by E loc = E0 + ∑ i [3(pi ⋅ ri )ri − ri2 pi ]/4πε 0 ri5 .
are distorted without generating dipole moments (acoustic mode). On the other hand, when
Figure 1.4c or d is stabilized, dipole moments are generated (optical mode). The final stabilized
states Figure 1.4c and d correspond to ferroelectric and antiferroelectric states, respectively. If
this particular mode becomes stabilized, with decreasing temperature, the vibration mode fre-
quency decreases (soft phonon mode), and finally at a certain phase transition temperature this
frequency becomes zero.
It follows that, at any individual A ion site, there exists a local field from the surrounding polar-
ization P, even if there is no external field. The concept of the local field is shown schematically in
Figure 1.5. It can be shown that
E loc = E0 + ∑ [3(p ⋅ r )r − r p ]/4πε r .

i
i i i i
2
i 0 i
5
= (γ /3ε 0 )P. (1.3)
This local field is the driving force for the ion shift. Here γ is called the Lorentz factor. For an
isotropic cubic system, it is known that γ = 1.1 ε0 is the permittivity of vacuum and is equal to 8.854 ×
10 −12 F/m. If the ionic polarizability of ion A is α, then the dipole moment of the unit cell of this
crystal is
μ = (αγ /3ε 0 )P. (1.4)
The energy of this dipole moment (dipole–dipole coupling) is
wdip = −μ ⋅ E loc = −(αγ 2/9ε 20 )P 2. (1.5)
Defining N to be the number of atoms per unit volume:
Wdip = Nwdip = −( N αγ 2/9ε 02 )P 2. (1.6)
Furthermore, when the A ions are displaced from their nonpolar equilibrium positions, the elastic
energy also increases. If the displacement is u, and the force constants are k and k¢, then the increase
of the elastic energy per unit volume can be expressed as
Welas = N [(k /2)u2 + (k ′ /4)u 4 ]. (1.7)

Wdip Welas
P
(a) (b)
Wtot
P
(c)
FIGURE 1.6 Energy explanation of the origin of spontaneous polarization. (a) Dipole interaction:
Wdip = – (Nag2/9e 02 )P2. (b) Elastic energy: Welas = (k ¢/2Nq2 )P2 + (k/4N3 q4 )P4. (c) Total energy: Wtot = Wdip + Welas.
Here, k¢ (>0) is the higher order force constant. It should be noted that in pyroelectrics, k¢ plays
an important role in determining the magnitude of the dipole moment. By rewriting Equation 1.7
using
P = Nqu, (1.8)
where q is the electric charge, and combining with Equation 1.6, the total energy can be expressed
as follows (see Figure 1.6):
Wtot = Wdip + Welas
= [(k /2Nq 2 ) − (N αγ 2 / 9ε 20 )] P 2 + [k ′ /4N 3q 4 ]P 4 . (1.9)
From this, one can see that if the coefficient of the harmonic term of the elastic energy is equal to or
greater than the coefficient of the dipole–dipole coupling, then P = 0; the A ions are stable and remain at
the nonpolar equilibrium positions. Otherwise, a shift from the equilibrium position (P2 = [(2Nαγ 2/9ε02)
− (k/Nq2)]/[k¢/N 3q4]) is stable. Spontaneous polarization can occur more easily in perovskite type crys-
tal structure (e.g., barium titanate, BT) due to a higher value of Lorenz factor γ (=10)2 than that of other
crystal structures. Note also that the polarizability is sensitive to temperature, leading to the phase tran-
sition. Suppose that the ionic polarizability of ion A, α, increases with decreasing temperature, even if
[(k/2Nq2) − (Nαγ 2/9ε02)] > 0 (paraelectric) at a high temperature, this value may become negative with
decreasing temperature, leading to a ferroelectric phase transition. Considering a first approximation,
a linear relation of the α with temperature, that is, the well-known Curie–Weiss law:
[(k /2Nq 2 ) − (N αγ 2 /9ε 20 )] ∝ (T − T0 )/C (1.10)
can be derived. This equation will be discussed again in detail in Section 2.2.2.
EXAMPLE PROBLEM 1.1

BaTiO3 exhibits ionic displacements as illustrated in Figure 1.7 at room temperature (RT).
Calculate the magnitude of the spontaneous polarization. The lattice constants are c = 4.036 Å
and a = 3.992 Å.
Hint
P = Nμ (N: number of the dipole moments included in a unit volume)

0.061 å
Ti4+ O2–
Ba2+
Ba2+ 0.036 å
Ti4+ 0.12 å
O2–
FIGURE 1.7 Ionic displacements in BaTiO3.
Solution
The dipole moment is calculated by taking the product of the charge magnitude and its
displacement. The total dipole moment in a unit cell is calculated by summing the Ba2+, Ti4+,
O2− related dipoles (notice the fractional contribution of each atom, i.e., 1/8 for corner atoms and
1/2 for face-centered atoms)
p = 8(2e /8)(0.061 × 10 −10 m) + 4e(0.12 × 10 −10 m) + 2(− 2e /2) (− 0.036 × 10 −10 m)
= 0.674 × 10 −10 e m
_ 29
= 1.08 × 10 C m, (P1.1.1)
where e = 1.602 × 10 −19 C. The unit cell volume is given by
v = a 2 c = (3.992)2 (4.036) × 10 −30 m 3
= 64.3 × 10 −30 m 3 . (P1.1.2)
The spontaneous polarization is defined as the polarization (total dipole moment) per unit
volume:
PS = p /v = 1.08 × 10 −29 C m / 64.3 × 10 −30 m 3
= 0.17C/m 2. (P1.1.3)
The experimental value of PS is about 0.25 C/m2.
1.3 ORIGIN OF FIELD-INDUCED STRAIN

Solids, especially ceramics (inorganic materials), are relatively hard mechanically, but still expand
or contract depending on the change of the state parameters. The strain (defined as the displacement
ΔL/initial length L) caused by temperature change and stress are known as thermal expansion and
elastic deformation, respectively. In insulating materials, the application of an electric field can also
cause deformation. This is called electric-field-induced strain.

Lattice constant Lattice constant

E E
– + – + – + – +
Ion pair potential energy Ion pair potential energy
δ1 δ2 δ1 δ2
(a) (b)
FIGURE 1.8 Microscopic explanation of the (a) piezoelectric strain and (b) electrostriction.
Generally speaking, the word electrostriction is used in a general sense to describe elec-
tric-field-induced strain, and hence frequently also implies the “converse piezoelectric effect.”
However, in solid-state theory, the converse piezoelectric effect is defined as a primary elec-
tromechanical coupling effect, that is, the strain is proportional to the electric field, while elec-
trostriction is a secondary coupling in which the strain is proportional to the square of the electric
field. Thus, strictly speaking, they should be distinguished. However, the piezoelectricity of a
ferroelectric, which has a centrosymmetric prototype (high temperature) phase, is considered to
originate from the electrostrictive interaction, and hence the two effects are related. The above
phenomena hold strictly under the assumptions that the object material is a monodomain single
crystal and that its state does not change under the application of an electric field. In a practi-
cal piezoelectric ceramic, additional strains accompanied by the reorientation of ferroelectric
domains are also important.
Why a strain is induced by an electric field is explained herewith.3 For simplicity, let us consider
an ionic crystal such as NaCl. Figure 1.8a and b shows a one-dimensional (1-D) rigid-ion spring
model of the crystal lattice. The springs represent equivalently the cohesive force resulting from
the electrostatic Coulomb energy and the quantum mechanical repulsive energy. Figure 1.8b shows
the centrosymmetric case, whereas Figure 1.8a shows the more general noncentrosymmetric case.
In Figure 1.8b, the springs joining the ions are all the same, whereas in Figure 1.8a, the springs
joining the ions are different for the longer and shorter ionic distances; in other words, hard
and soft springs existing alternately are important. Next, consider the state of the crystal lattice
Figure 1.8a under an applied electric field. The cations are drawn in the direction of the electric
field and the anions in the opposite direction, leading to the relative change in the interionic
distance. Depending on the direction of the electric field, the soft spring expands or contracts
more than the contraction or expansion of the hard spring, causing a strain x (a unit cell length
change) in proportion to the electric field E. This is the converse piezoelectric effect. When
expressed as
x = dE, (1.11)
the proportionality constant d is called the piezoelectric constant.

On the other hand, in Figure 1.8b, the amounts of extension and contraction of the spring
are nearly the same, and the distance between the two cations (lattice parameter) remains
almost the same, hence, there is no strain. However, more precisely, ions are not connected by
such idealized springs (those are called harmonic springs, in which force (F) = spring constant

(k) × displacement (Δ) holds). In most cases, the springs possess anharmonicity (F = k1Δ − k2Δ2),
that is, they are somewhat easy to extend but hard to contract. Such subtle differences in the dis-
placement causes a change in the lattice parameter, producing a strain which is independent of the
direction of the applied electric field (+E or −E), and hence is an even function of the electric field.
This is called the electrostrictive effect and can be expressed as
x = ME 2 , (1.12)
where M is the electrostrictive constant.

The 1-D crystal pictured in Figure 1.8a also possesses a spontaneous bias of electrical charge, or
a spontaneous dipole moment. The total dipole moment per unit volume is called the spontaneous
polarization. When a large reverse bias electric field is applied to a crystal that has a spontaneous
polarization in a particular polar direction, a transition “phase” is formed which is another stable
crystal state in which the relative positions of the ions are reversed. (In terms of an untwinned
single crystal, this is equivalent to rotating the crystal 180° about an axis perpendicular to its polar
axis. This is also understood from the potential double minima in Figure 1.6.) This transition,
referred to as polarization reversal, also causes a remarkable change in strain. This particular class
of substances is referred to as ferroelectrics, as mentioned in Section 1.1. Generally, what is actu-
ally observed as a field-induced strain is a complicated combination of the three basic effects just
described.
Figure 1.9 shows typical strain curves for a piezoelectric lead zirconate titanate (PZT)-based
ceramic and an electrostrictive lead magnesium niobate (PMN)-based ceramic.4 An almost linear
strain curve in PZT becomes distorted and shows large hysteresis on increasing applied electric
field level, which is due to the polarization reorientation. On the other hand, PMN does not exhibit
hysteresis under an electric field cycle. However, the strain curve deviates from the quadratic rela-
tion (E2) at a high electric field level.
We described the converse piezoelectric effect above. Then, what is the normal or direct piezo-
electric effect? This is the phenomenon whereby charge (Coulomb per unit area) is generated by
applying an external stress (force per unit area). Note that the same piezoelectric coefficient d in
Equation 1.11 is used in the relation
P = dX . (1.13)
Strain ΔI/I
×10–3 ×10–3
4 1
Strain ΔI/I
3 0.75
2 0.5
1 0.25
–15 –10 –5 0 5 10 15 –15 –10 –5 0 5 10 15

(a) Electric field (kV/cm) (b) Electric field (kV/cm)
FIGURE 1.9 Typical strain curves for a piezoelectric PZT-based ceramic (a) and an electrostrictive
PMN-based ceramic (b).

EXAMPLE PROBLEM 1.2

One of the PZT ceramics has a piezoelectric constant of d33 = 590 × 10 −12 C/N with a dielectric
constant ε3 = 3400 and an elastic compliance s33 = 20 × 10 −12 m2/N.
(a) Calculate the induced strain under an applied electric field of E3 = 10 × 105 V/m. Then,
calculate the generative stress under a completely clamped condition.
(b) Calculate the induced electric field under an applied stress of X3 = 3 × 107 N/m2. The
induced field does not coincide with the above-mentioned 10 × 105 V/m. Explain the rea-
son, taking account of electromechanical coupling factor k.
Hint
We will introduce the tensor quantities in Chapter 2. However, try to solve the problems by
neglecting the subscripts at the moment, in order to make the reader understand the concept of
electromechanical coupling factor k.
x = dE, (P1.2.1)
P = dX , (P1.2.2)
and
k 2 = d 2/(sε 0 ε). (P1.2.3)
Solution
_
(a) x3 = d3 E3 = (590 × 10 12
C/N)(10 × 10 5 V/m)
_
= 5.9 × 10 4. (P1.2.4)
Under a completely clamped condition
_ _12
X3 = x3 /s33 = 5.9 × 10 4/ 20 × 10 m 2 /N
= 3.0 × 10 7 N/m 2. (P1.2.5)
(b) P3 = d33 X3 = (590 × 10 −12 C /N) (3 × 10 7 N/m 2 )
= 1.77 × 10 −2 C/m 2. (P1.2.6)
_
E3 = P3 /ε 0 ε 3 = (1.77 × 10 −2 C/m 2 )/(3400 × 8.854 × 10 12
F/m)
= 5.9 × 10 5 V/m. (P1.2.7)
The induced field is only 59% of the field in the case of (a); this is explainable by the electrome-
chanical coupling factor k. When electric energy is supplied to a piezoelectric sample and some
part is transduced into mechanical energy, the electromechanical coupling factor k is defined by
k 2 = (Stored mechanical energy)/(Input electrical energy).
= (1/2)(x 2/s)/(1/2)(ε 0 εE 2 ) = (1/2)((dE )2/s)/(1/2)(ε 0 εE 2 )
= d 2/sε 0 ε. (P1.2.8)

On the other hand, when mechanical energy is supplied to the sample and some part is trans-
duced into electrical energy, k is also defined (notice the same result) by
k 2 = (Stored electrical energy)/(Input mechanical energy).
= (1/2)(P 2/ε 0 ε)/(1/2)sX 2 = (1/2)((dX )2 /ε 0 ε)/(1/2)sX 2
= d 2 /sε 0 ε. (P1.2.9)
In the above case
2
k33 = d33
2
/s33ε 0 ε 3
= (590 × 10 −12 C/N)2 /(20 × 10 −12 m 2 /N)(3400 × 8.854 × 10 −12 F/m)
= 0.58. (P1.2.10)
The reason why the induced field in (b) is about k2 of the 10 × 105 V/m of (a) is this transduction
ratio; that is, each energy transduction ratio accompanying the E → M and M → E processes
is k2. Thus (last electrical energy)/(initial electrical energy) = k 4, leading to (last induced field)/
(initial applied field) = k2.
1.4 ELECTRO-OPTIC EFFECT

Since light is an alternating electromagnetic wave with electric and magnetic field vibration direc-
tions mutually perpendicular to one another, the electric field induces an electric polarization in a
dielectric crystal and the light itself is influenced by the crystal. The alternating frequency of the
light is so high (PHz = 1015 Hz) that only the electronic polarization can follow the electric field
change (see Figure 1.2), and the relative permittivity of an optically transparent crystal is small, not
exceeding 10. The relative permittivity ε at this high frequency is related to the refractive index n
by the following equation:
ε = n2 . (1.14)
When an external electric field is applied to the crystal, ion displacement is induced, deforming the
shape of the electron cloud, and consequently the refractive index is changed. This phenomenon is
called the electro-optic effect.
Generally, the refractive index is treated as a symmetrical second-rank tensor quantity and is
represented geometrically by the optical indicatrix which is described by
x 2 y2 z2
2
+ 2 + 2 = 1, (1.15)
n1 n2 n3
where n1, n2, and n3 are the principal refractive indices. With the application of an electric field, the
change in refractive index is given by a Taylor expansion expression in terms of E:
1/nij2 (E ) − 1/nij2 (0) = ∑r E + ∑ R

ijk k ijkl Ek El . (1.16)
Here n(E) and n(0) are the refractive indices at E and zero field, respectively, and rijk is the pri-
mary electro-optic coefficient (Pockels effect) and Rijkl is the secondary coefficient (Kerr effect).
Remember that the expansion is not based on nij, but on 1/n2ij.

Considering the paraelectric phase of a perovskite crystal (m3 m) as an example, the Kerr coef-
ficients are represented in the following matrix:
⎛ R11 R12 R12 0 0 0 ⎞

⎜ ⎟
⎜ R12 R11 R12 0 0 0 ⎟
⎜ R12 R12 R11 0 0 0 ⎟
⎜ ⎟
⎜ 0 0 0 R44 0 0 ⎟
⎜ 0 0 0 0 R44 0 ⎟
⎜⎜ ⎟
⎝ 0 0 0 0 0 R44 ⎟⎠
so that the refractive indicatrix under an electric field applied along the z-direction is expressed as
x 2 + y2 z2
2
+ 2
= 1. (1.17)
⎡ ⎛ n2 ⎞ ⎤ ⎡ ⎛ n2 ⎞ ⎤
n ⎢1 − ⎜ 0 ⎟ R12 Ez2 ⎥
2
0 n ⎢1 − ⎜ 0 ⎟ R11Ez2 ⎥
2
0
⎣⎢ ⎝ 2 ⎠ ⎦⎥ ⎣⎢ ⎝ 2 ⎠ ⎦⎥
The refractive index change under an external electric field is explained intuitively as follows. When
an electric field Ez is applied to a cubic perovskite crystal, the crystal is elongated along the z-axis
and contracted along both the x- and y-axes. Consequently, the material’s density or compactness
will be decreased along the z-axis and densified along the x- and y-axes, leading to a decrease in
the refractive index nz and an increase of the indices nx and ny, as shown in Figure 1.10 (sphere
becomes a doughnut-shape). Note that the refractive index is proportional to the electron density
or ion compactness along the polarized light electric field direction which is perpendicular to the
light propagation direction. Taking into account the above description, R11 and R12 in Equation 1.16
should be positive and negative, respectively, in this perovskite structure.
When light is transmitted along the y-direction, the phase retardation Γy between the ordinary
(polarized along the x orientation) and extraordinary waves (polarized along the z orientation) is
given by (refer to Figure 1.11)
Γ y = (2π /λ)(n03 /2)(R11 − R12 )L (Vz /d )2 , (1.18)
z
nz
y
ny
nx
x
FIGURE 1.10 Refractive indicatrix change of a cubit crystal with electric field (Kerr effect) (left). An origi-
nal sphere becomes a doughnut shape. Perovskite unit cell change with electric field (right).

Unpolarized light
+ L
45° ne
n0 d
Polarizer
–45°
Electro-optic crystal
Polarizer
FIGURE 1.11 Optical phase retardation through an electro-optic crystal. Notice the crossed polarizer
configuration.
where
d is the electrode gap
L is the optical path length
Placing the crystal between crossed polarizers arranged at an angle of 45° (one +45° and the other
−45°) with respect to the z-axis, the output light intensity is modulated as a function of applied
voltage according to
I = I 0 sin 2 (Γ y /2) = (1/2)I 0 (1 − cos Γ y ). (1.19)
This is the principle behind the operation of a light shutter/valve, and the voltage required for the
first intensity maximum (i.e., Γy = π) is an important characteristic called the half-wave voltage. The
detailed derivation process is given in Example Problem 1.3.
EXAMPLE PROBLEM 1.3

A paraelectric La-doped PZT (PLZT) sample with a rectangular shape (optical pathlength: L)
is put between two crossed polarizers arranged at 45° with respect to the z-axis (Figure 1.11).
Assuming a voltage Vz is applied to the sample with an electrode gap d, and light with a wave-
length of λ (incident intensity: I0) is transmitted through it, derive the output light intensity I(Vz)
by neglecting light absorption in the PLZT. Other parameters are listed below:
Refractive index at E = 0: n 0
Electro-optic Kerr coefficients: R11, R12
Phase retardation: Γy
Reflectance at the crystal surface (normal incidence): Re = [(n − 1)/(n + 1)]2
Hint
The initial spherical indicatrix will be deformed into an ellipsoidal one under an applied electric
field Ez:
x 2 + y2 z2
2
+ 2
= 1. (P1.3.1)
⎡ ⎛ n2 ⎞ ⎤ ⎡ ⎛ n2 ⎞ ⎤
n ⎢1 − ⎜ 0 ⎟ R12 E2z ⎥
2
0 n ⎢1 − ⎜ 0 ⎟ R11E2z ⎥
2
0
⎣⎢ ⎝ 2 ⎠ ⎥⎦ ⎣⎢ ⎝ 2 ⎠ ⎥⎦
The output light intensity is reduced twice, once at the inlet and once again at the outlet crystal
surfaces by a factor of (1 − Re)2.

Also you should notice that the incident light (after passing through the first polarizer) has
ordinary and extraordinary light components of equal magnitudes, because of 45° cant angle of
the polarizer.
Solution
In a cubic structure, the refractive index change under an external electric field along z-axis is
expressed by the following two equations:
1/nz2 (Ez ) − 1/n02 = R11Ez2 , (P1.3.2)
1/nx2 (Ez ) − 1/n02 = R11Ez2 . (P1.3.3)
Taking into account the relation, d(1/n2) = −(2/n3)dn
nz (Ez ) = n0 −(1/2)n03 R11Ez2 , (P1.3.4)
nz ( Ez ) = n0 −(1/2)n03 R12 Ez2 , (P1.3.5)
where R11 > 0 and R12 < 0 in most cases.

Since the wavelengths of the extraordinary (polarized along z-direction) and the ordinary
(polarized along x) waves are described as
λ z = λ 0 /nz , (P1.3.6)
λ x = λ 0 /nx , (P1.3.7)
where λ0 is the vacuum wavelength of the incident light, and the waves existing in the crystal
with an optical pathlength of L are L/λz and L/λx, respectively, the phase difference between
these waves (retardation Γy) is given by
Γ y = 2π(L /λ x − L /λ z ) = (2π /λ 0 )L Δn
= (2π /λ 0 )L (1/2)n03 (R11 − R12 )(Vz /d )2 . (P1.3.8)
The retardation is proportional to the birefringence Δn (=nx – nz).

First, in comparison with the unpolarized light intensity, the intensity from one polarizer is
just a half (by neglecting the light absorption by the polarizer). We will take this intensity from
one polarizer as the incident light intensity I0. Second, we describe the linearly polarized light
incident on the PLZT in terms of its electric field vector as
⎛ ⎡⎛ 2π ⎞ ⎤⎞
⎜ sin ⎢⎜ ⎟ y − ωt ⎥⎟
⎛ ex ⎞ ⎜ ⎣⎝ λ 0 ⎠ ⎦⎟
⎜⎝ e ⎟⎠ = l0 ⎜ . (P1.3.9)
⎡⎛ 2π ⎞ ⎤⎟
z
⎜ sin ⎢ ⎟
⎜⎝ ⎜ ⎟ y − ωt ⎥ ⎟
⎣⎝ λ 0 ⎠ ⎦⎠
Note here that the 45° arrangement generates the same electric field amplitude along x- and
z-directions, and that the relationship I = e2/2, i.e., the intensity is an average of e2 in terms of
time. Third, the output light from the PLZT can be described as
⎛ ⎡⎛ 2π ⎞ ⎤ ⎞
⎜ sin ⎢⎜ ⎟ y − ωt + φ⎥ ⎟
⎛ ex ⎞ ⎜ ⎣⎝ λ 0 ⎠ ⎦ ⎟
⎜⎝ e ⎟⎠ = (1 − Re ) l0 ⎜
2
. (P1.3.10)
⎡ ⎤⎟
z
⎜ sin ⎢⎛ 2π ⎞ y − ωt + φ − Γ y ⎥⎟
⎜⎝ ⎜ ⎟ ⎟
⎣⎝ λ 0 ⎠ ⎦⎠

Remember to consider two reflections on the front and backside of the PLZT sample. Also,
basically the extraordinary ray speed corresponding to output ez is faster than the ordinary ray
speed (ex), leading to the phase delay of Γy in ez when we consider the optical pathlength L.
Fourth, through the polarizer arranged at the −45° orientation, the electric field component in
this direction e−45cant is represented (taking into account the electric field vector projection on this
−45° orientation) by
l0
ex / 2 − ez / 2 = (1 − Re) {sin[(2π /λ 0 )y − ωt + φ] − sin[(2π /λ 0 )y − ωt + φ − Γ y ]}
2
l0
= (1 − Re) [(1 − cos Γ y ) sin[(2π /λ 0 )y − ωt + φ]
2
+ sin Γ y cos[(2π /λ 0 )y − ωt + φ]. (P1.3.11)
Thus, finally, the output intensity through the 2nd polarizer (−45° cant) is obtained from the
2
relation I = e−45cant /2:
I = (1/2)(1 − Re)2 (I 0 /2)[(1 − cos Γ y )2 + (sin Γ y )2 ]
⎛ 1 − cos Γ y ⎞
= (1 − Re)2 I 0 ⎜ ⎟⎠ .
(P1.3.12)
⎝ 2
Figure 1.12 shows the output intensity I as a function of applied voltage Vz, which exhibits maxi-
mum and minimum successively. The half-wave voltage, which is defined as the voltage required
to exhibit the first maximum in the transmitted light intensity, is given from the condition, cos Γy
= −1 or Γy = π. When we use a sample with Kerr (secondary) electro-optic effect, Γy is expressed
by Equation P1.3.8, and the half-wave voltage can be obtained as
λ0
Vz,λ /2 = d . (P1.3.13)
Ln03 ( R11 − R12 )
The intensity maxima can be sequentially obtained on increasing the applied voltage. Notice that
these half-wave, one-and-a-half-wave, two-and-a-half-wave voltages are not arranged linearly,
due to the nonlinear electro-optic effect.
Lab Demonstration 1.1—Polarizer
This is a demo with two polarizers by changing the angle between two. The reader is expected to
understand the polarized light characteristics (Figure 1.13).
Light
intensity
(1–Re)2 I0
0 Vz,λ/2 Applied voltage
FIGURE 1.12 Variation in the light intensity of a Kerr-effect electro-optic shutter with applied voltage.

(a) (b)
(c) (d)
FIGURE 1.13 Demo with a pair of optical polarizers by changing the angle. (a) 0°, (b) 30°, (c) 60°, and
(d) 90° angles.
(a) (b)
(c)
FIGURE 1.14 Demo with a pair of optical polarizers by changing the angle, sandwiching a Scotch tape.
(a) 0°, (b) 90°, and (c) 45° angles.
Lab Demonstration 1.2—Scotch Tape Magic
Scotch tape is different from other manufacturers’ products. A Scotch tape is sandwiched by a
pair of polarizers. Dark and bright conditions are reversed in comparison with Lab Demo 1.1,
as demonstrated in Figure 1.14a and b. When the analyzer (second polarizer) is arranged along
the tape extended direction, no interference is observed. This is the basis for understanding the
birefringence and retardation of the inserted optical material between a pair of polarizers.
Comments
Scotch tape is manufactured by stretching the tape, so that the refractive index along the tape is
different from that of the width direction. This birefringence and the optical pathlength (tape thick-
ness) generate accidentally the retardation Γy of π (refer to Equation P1.3.8). This situation exhibits
the above-mentioned “Scotch tape” magic. By inserting an electro-optic PLZT (replacing Scotch
tape) and inducing the birefringence with electric field, a similar dark and bright conversion can
be introduced in a PLZT light valve.

1.5 EXAMPLE OF FERROELECTRICS

Quartz (SiO2) and zinc oxide (ZnO) are popular piezoelectrics, but non ferroelectric (paraelec-
tric) materials. The “direct” piezoelectric effect was discovered in quartz by Pierre and Jacques
Curie in 1880. The “converse” piezoelectric effect was discovered successively in 1881 by Gabriel
Lippmann. The shipwreck of Titanic and World War I were the motivation for the undersea
transducer, sonar. Paul Langevin developed the so-called Langevin-type transducer, which was
originally composed of natural tiny quartz single crystals sandwiched by two metal blocks, in
order to tune the transducer resonance frequency around 26 kHz which is a desired range for
underwater applications.
On the other hand, the first ferroelectric discovered is Rochelle salt (NaKC4H4O6 . 4 H2O) in 1921.
Though this material was studied from an academic viewpoint, it has not been widely utilized in
practice because it is water-soluble and its Curie temperature is just above RT. KH2PO4 (KDP) was
the second discovery in 1935, which is also water-soluble and the Curie temperature is −150°C. We
had to wait until World War II for the third and most famous ceramic ferroelectric, i.e., BT (BaTiO3),
which was actually commercialized first as a transducer material. In order to develop compact
capacitors for portable “radar” systems to be used in the battle fields, TiO2-based conventional “con-
denser materials” were widely researched by doping various ions such as CaO, SrO, BaO, MgO, and
Fe2O3. Three groups in the United States, Russia, and Japan discovered BaTiO3 almost the same
time as the World War II.
1.5.1 BARIUM TITANATE

We describe basic properties of ferroelectrics using BT as an example. As shown in Figure 1.15,
BaTiO3 has a perovskite crystal structure. In the high-temperature paraelectric phase (nonpolar
phase) there is no spontaneous polarization (the symmetry is Oh – m3 m). Below the transition tem-
perature TC called the Curie temperature (about 130°C), spontaneous polarization occurs, and the
crystal structure becomes slightly elongated, that is, tetragonal (C4 v – 4 mm). The detailed ionic
shift is illustrated in Figure 1.7. Figure 1.16 shows schematically the temperature dependence of the
spontaneous polarization PS and permittivity (dielectric constant) ε in a BT-like first-order transition
ferroelectric. PS decreases with increasing temperature and vanishes at the Curie temperature, while
ε tends to diverge near TC. Also, the reciprocal (relative) permittivity 1/ε is known to be linear with
respect to the temperature over a wide range in the paraelectric phase (so-called Curie–Weiss law)
ε = C /(T − T0 ), (1.20)
Ti4+
Ba2+
O2–
T > TC T < TC
TC: Curie temperature
FIGURE 1.15 Crystal structures of BaTiO3: Higher than TC (left) and lower than TC (right).

Physical properties (a) Capacitor
Permittivity ε
Spontaneous PS
(b) Memory Inverse permittivity 1/ε
(c) Pyrosensor
T0 TC
Temperature
(Curie temp.)
(e) Piezoelectric transducer (d) Electro-optic device

(f) Electrostatrictor
FIGURE 1.16 Temperature dependence of the spontaneous polarization and permittivity in a ferroelectric
material. (a)–(f) indicate the temperature ranges for each application.
where
C is the Curie–Weiss constant
T0 is the Curie–Weiss temperature
T0 is slightly lower than the exact transition temperature TC.

It is also known that the spontaneous polarization PS and the spontaneous strain xS follow the
relationship:
xS = QPS2 , (1.21)
and xS decreases almost linearly with increasing temperature. In the case of BaTiO3, it exhibits the
piezoelectric effect in the ferroelectric phase, while in the paraelectric phase, it is nonpiezoelectric
and exhibits only the electrostrictive effect.
If the above consideration is rephrased from the application viewpoints, the Curie temperature
for capacitor materials is designed to set around RT; TC for memory materials around 100°C higher
than RT; TC for pyrosensors just above RT; TC for piezoelectric transducers typically much higher
than RT, higher than 200°C; TC for electro-optic and electrostrictive devices are lower than RT to
use their paraelectric state. In other words, we will design practical materials with their Curie points
suitable for each application.
With decreasing temperature from RT, however, BT undergoes a series of complicated phase
transitions. Figure 1.17 illustrates these successive phase transitions.
1.5.2 PZT
PZT solid solution systems were discovered in 1954 by Japanese researchers, Gen Shirane, Etsuro
Sawaguchi et al. However, the enormous piezoelectric properties were discovered by B. Jaffe, Clevite
Corporation, and Clevite took the most important PZT patent for transducer applications. Because

Rhombohedral Orthorhombic Tetragonal Cubic
Ps Ps Ps
Dielectric constant
10,000
5,000 εa
εc
0
–150 –100 –50 0 50 100 150
Temperature (°C)
FIGURE 1.17 Various phase transitions in BT.
of this strong basic patent, Japanese ceramic companies were encouraged actually to develop ter-
nary systems to overcome the performance, and more importantly, to escape from the Clevite’s
patent; that is, PZT + a complex perovskite such as Pb(Mg1/3Nb2/3)O3 (Matsushita Panasonic),
Pb(Ni1/3Nb2/3)O3 (NEC), and Pb(Zn1/3Nb2/3)O3 (Toshiba), which are the basic compositions in recent
years. The performance details are discussed in Chapter 7.
1.5.3 POLYVINYLIDENE DIFLUORIDE (PVDF)

Thanks to Kawai’s efforts, polyvinylidene difluoride (PVDF or PVF2) was discovered in 1969.
Though the piezoelectric d constant is not as high as piezoceramics, high piezoelectric g constant
due to small permittivity is attractive from the sensor application viewpoint.
PVDF is a polymer with monomers of CH2CF2. Since H and F have positive and negative ion-
ization tendency, the monomer itself has a dipole moment. Crystallization from the melt forms the
nonpolar α-phase, which can be converted into the polar β-phase by a uniaxial or biaxial drawing
operation; the resulting dipoles are then reoriented through electric poling (see Figure 1.18).
Large sheets can be manufactured and thermally formed into complex shapes. Piezoelectric
polymers have the following characteristics: (a) small piezoelectric d constants (for actuators) and
large g constants (for sensors), (b) light weight and soft elasticity, leading to good acoustic imped-
ance matching with water or the human body, (c) a low mechanical quality factor Qm, allowing for
a broad resonance band width.
(CH2CF2)n
Carbon
Fluoride
z Hydrogen
FIGURE 1.18 Structure of PVDF.

1.5.4 PB-FREE PIEZOCERAMICS

In 2006, the European Community started restrictions on the use of certain hazardous substances
(RoHS), which explicitly limits the usage of lead (Pb) in electronic equipments. Basically, we may the
need to regulate the usage of PZT, the most widely used piezoelectric ceramics, in the future. Japanese
and European communities may experience governmental regulation on the PZT usage in these 10 years.
Pb-free piezoceramics have started to be developed after 1999. The Pb-free materials include (1) (K,Na)
(Ta,Nb)O3-based, (2) (Bi,Na)TiO3, and (3) BaTiO3, which reminds us that “history will repeat.”
1.6 APPLICATIONS OF FERROELECTRICS

Ferroelectric materials, especially polycrystalline ceramics, are very promising for a variety of
applications such as high-permittivity capacitors (Chapter 4), ferroelectric memories (Chapter 5),
pyroelectric sensors (Chapter 6), piezoelectric and electrostrictive transducers (Chapter 7), electro-
optic devices (Chapter 8), and PTC thermistors (Chapter 9).
For capacitor dielectrics, the peak dielectric constant around the transition temperature (Curie) is
utilized, while for memory applications, the material must be ferroelectric at RT (refer to Figure 1.16).
A large temperature dependence of the spontaneous polarization below TC is sought for pyroelectric
sensors. The converse pyroelectric effect is the electrocaloric effect (electric field generates the tem-
perature decrease), which is becoming a hot research topic in this energy-saving age.
Piezoelectric materials are used as sensors and actuators, where the TC should be much above
RT. Pressure and acceleration sensors are now commercially available in addition to conventional
piezovibrators. Precision positioners and pulse drive linear motors have already been installed in pre-
cision lathe machines, semiconductor manufacturing apparatuses, and office equipment. Recently,
enthusiastic development is found in ultrasonic motors, aiming at electromagnetic, noise-free, and
very compact motor applications. In parallel to the new energy source programs, piezoelectric
energy harvesting systems have become popular. Waste mechanical energy such as noise vibration,
wind, and human walking can be converted into electrical energy and can be used directly for signal
transmission or to charge up batteries for portable electronics.
Electro-optic materials will become key components in displays and optical communication
systems in the future. Optical beam scanners, light valves, and switches are urgent necessities. For
thermistor applications, semiconductive ferroelectric ceramics with a PTC of resistivity based on a
junction effect have been developed from BT-based materials. We will discuss the above applica-
tions in detail in the following chapters.
Lab Demonstration 1.3—Applications in Cellular Phones
Portable communication devices such as cordless, portable, and car telephones have become
popular worldwide. The reader is requested to identify what kind of ferroelectric and dielectric
components are used in a cellular phone. An old design phone is dismantled to see reasonably
large components (Figure 1.19).
Comments
Chip monolithic ceramic capacitors
Microwave oscillators
Microwave filters
Chip monolithic LC filters
Ceramic resonators
High-frequency surface acoustic wave (SAW) filters
Ceramic filters
Piezoelectric receivers
Piezoelectric speakers
Recent designs include ferroelectric memory, piezoactuators for camera autozoom/focus
mechanisms.

FIGURE 1.19 Dismantled cellular phone. Identify the ferroelectric components.
Chapter Essentials
1. Category:
Dielectrics > Piezoelectrics > Pyroelectrics > Ferroelectrics
2. Ferroelectric materials are very promising for a variety of applications:
High-permittivity capacitors
Ferroelectric memories
Pyroelectric sensors, electrocaloric fridges
Piezoelectric/electrostrictive transducers, piezoelectric energy harvesting
Electro-optic devices
PTC thermistors
3. Origin of spontaneous polarization:
Dipole coupling with the local field—driving force of ionic displacement
Elastic anharmonic term—impeding ionic displacement
4. Field-induced strains:
Piezoelectric strain—x = dE in an asymmetric crystal
Strain associated with polarization reorientations
Electrostriction—x = ME2 in a symmetric crystal
Piezostriction: the difference of the harmonic term of the two equivalent springs
Electrostriction: anharmonicity of the equivalent springs
5. Electro-optic effect: refractive index change with an external electric field. The secondary
Kerr effect is frequently used. Devices making use of this effect require a pair of crossed
polarizers arranged at a 45° angle with respect to the electric field direction.
Half-wave voltage: minimum voltage required for an electro-optic crystal to exhibit the first
maximum light intensity.
Check Point
1. There are three microscopic origins of polarization: electronic polarization, ionic polariza-
tion, and [— (a)]. Fill in the blank.
2. The local field is the driving force for the spontaneous polarization. How is γ called, which
enhances the applied electric field E.

3. (T/F) Taking into account a famous relation between the relative permittivity ε and the
refractive index n: ε = n2, we can obtain n = 4 for the crystal with ε = 16 (at 1 kHz). Is this
argument true or false?
4. The randomly oriented light passed through a polarizer. When we neglect the light absorp-
tion by the polarizer, what percentage of the light intensity can we obtain from the polar-
izer in comparison with the input intensity?
5. (T/F) The definition of Pockels electro-optic coefficient r1jk is given by an expansion expres-
sion: 1/nij ( E ) − 1/nij (0) = Σrijk Ek . True or False ?
6. (T/F) The Curie temperature of a capacitor ferroelectric material should be around RT.
True or False?
7. (T/F) The Curie temperature of a piezoelectric transducer material should be 100°C lower
than RT. True or False?
8. (T/F) The Curie temperature of a pyroelectric sensor material should be 200°C higher than
RT. True or False?
9. Provide a name of a representative polymer piezoelectric.
10. Provide a full expression of PZT.
Chapter Problems
1.1 Consider a 1-D infinite chain of two kinds of ions +q and −q, arranged alternately over
a distance a. Calculate the Coulombic potential energy at one +q ion, and obtain the
Madelung energy (Figure 1.20).
Hint
Remember U = −(1/4πε 0 ε)(q 2/r ).
Use an expansion series of log(1 + x) = x – x2/2 + x3/3 – x4/4 + …

This is the basic concept of a crystal structure stability, which will be discussed again
in Chapter 3 for considering nanoparticle ferroelectrics.
1.2 Potassium tantalate niobate K(Ta0.65Nb0.35)O3 is cubic at RT. By applying an electric

field Ez along a perovskite [100] axis, it exhibits an induced polarization P3, electrostric-
tive strains x3 = Q11P 32 and x1 = Q12P 32, and refractive index changes Δn3 = −(1/2)n 03g11P 32
and Δn1 = −(1/2)n 03g12P 32. Experimental values of these are Q11 = 0.090 m4/C2, Q12 =
−0.035 m4/C2; g11 = 0.136 m4/C2, g12 = −0.038 m4/C2. Comparing the absolute values
between Q and g and the ratios Q11/Q12 and g11/g12, discuss similarities in terms of the
crystal lattice compactness along and perpendicular to the electric field.
1.3 The following data was cited from Intel Corporation’s famous “Gordon Moore’s Law.”5
The number of transistors per die will increase logarithmically with year. From this
expectation, the following technologies will be definitely required (Figure 1.21):
(1) 90 nm lithography technology
(2) Replacement of the present gate oxide dielectric with a high-k oxide
+q –q
–4a –3a –2a –a 0 +a +2a +3a +4a
FIGURE 1.20 1-D atomic model.

Integrate-circuit complexity
Transistors
per die
1010 1965 Actual data 4G
1G2G
109 MOS Arrays MOS Logic 1975 actual data 512M
256M
128M ItaniumTM
108 1975 Projection 64M Pentium4â
Memory 16M PentiumIIIâ
107 4M
Microprocessor 1M PentiumIIâ
106 256K Pentiumâ
1486TM
64K
105 1386TM
4K 80286
104 16K 8086
1K
8008
103 4004
102
101
100
1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010
FIGURE 1.21 Gordon Moore’s Law. (Courtesy of Intel Corporation, Santa Clara, CA, http://www.intel.com/
technology/mooreslaw.)
Questions
(a) In order to realize the above requirements (1) and (2), what sort of ferroelectric
devices should be developed? Consider these “Needs-pull” development strategies.
(b) What other ferroelectric applications can be considered, relating to “integrated-
circuit complexity?”
1.4 Expanding Lab Demo 1.3, collect more detailed information on practical designs of
multilayer capacitors (how small size?), SAW filters (what frequency?), speakers and
microphones (operation frequency range difference among speakers and microphones),
memories (flush or ferroelectric memories?), and piezoactuators in the recent cel-
lular phones. Note that the development of phone camera modules is remarkable in
particular.
REFERENCES
1. C. Kittel: Introduction to Solid State Physics, 6th edn., Chap. 13, John Wiley & Sons, New York (1986).
2. W. Kinase, Y. Uemura and M. Kikuchi: J. Phys. Chem. Solids, 30, 441 (1969).
3. K. Uchino and S. Nomura: Bull. Jpn. Appl. Phys., 52, 575 (1983); K. Uchino, S. Nomura, L. E. Cross, R.
E. Newnham and S. J. Jang: Electrostrictive effect in perovskites and its transducer applications, J. Mater.
Sci., 16, 569 (1981).
4. K. Uchino: Electrostrictive actuators: materials and applications, Bull. Am. Ceram. Soc., 65(4), 647
(1986).
5. Intel Corporation, Santa Clara, CA, http://www.intel.com/technology/mooreslaw/

2 Mathematical Treatment
of Ferroelectrics
Physicists usually treat a natural phenomenon using a simple mathematical form: one is a linear
approximation and another is a nonlinear expansion theory. Hooke’s law, the stress–strain relation
and Ohm’s law, the voltage–current relation are two of the most famous linear laws in physics.
These linear relations are extended into matrix or tensor relations in linear algebra. We will discuss
the piezoelectric tensor/matrix in Section 2.1. On the other hand, the Maclaurin or Taylor expansion
series are popularly used to calculate slightly perturbed physical quantities around an equilibrium
state including nonlinear effects. We will consider phenomenology of ferroelectricity (nonlinear
relation) and antiferroelectricity in Sections 2.2 and 2.3, respectively. Section 2.4 treats the ferro-
electric domain contributions to the physical properties.
“If the facts don’t fit the theory, change the facts!” This is a famous quote by Dr. Albert Einstein.
The author of this textbook is not so arrogant, but his stand point is that “Theory corresponds to a
blind person’s cane.” Knowing the theory helps in reducing unnecessary effort in the experiment,
however, the reader is never recommended to change or to create fraud data on the experimental
result with a strongly biased thought.
2.1 TENSOR REPRESENTATION OF PHYSICAL PROPERTIES

2.1.1 FUNDAMENTALS OF TENSOR
When we correlate one input physical parameter X to output parameter Y, we introduce a propor-
tional parameter A in a linear relation approximation:
Y = AX . (2.1)
When X is a p-rank tensor, and Y is a q-rank tensor, A is supposed to be expressed by using a (p + q)-
rank tensor:
Yij…q = ∑A ij…qlm…p Xlm…p . (2.2)
A p-rank tensor Xlm…p is represented based on an (x, y, z) coordinate system. When this tensor is
described in an (x′, y′, z′) coordinate system, a new tensor X′ can be represented as
Xij′ … p = ∑a a
lm − p
il jm … a pq Xlm…q , (2.3)
where aij is called transformation matrix, which is a unitary matrix without an imaginary part,
satisfying the following relation (inverse matrix º transposed matrix):
−1
⎛ a11 a12 a13 ⎞ ⎛ a11 a21 a31 ⎞
⎜a a22 a23 ⎟ = ⎜ a12 a22 a32 ⎟ . (2.4)
⎜ 21 ⎟ ⎜ ⎟
⎝ a31 a32 a33 ⎠ ⎝ a13 a23 a33 ⎠
25

For centrosymmetry, since x → −x, y → −y, z → −z, the transformation matrix is written as
⎛ −1 0 0⎞
⎜ ⎟
⎜0 −1 0⎟ (2.5)
⎜0 −1 ⎟⎠
⎝ 0
and for rotation (angle θ) about a principal z-axis
⎛ cos θ sin θ 0⎞
⎜ − sin θ cos θ 0⎟ . (2.6)
⎜ ⎟
⎝ 0 0 1⎠
2.1.2 TENSOR REPRESENTATION

Let us first consider the tensor for electric conductivity. The conductivity is defined so as to correlate
an applied electric field E and the induced current density J as follows1:
J = σE . (2.7)
Since both the electric field and the current density are first-rank tensor (i.e., vector) quantities,
the conductivity should have a second-rank tensor representation (i.e., with two suffixes); this is
described as
⎛ J1 ⎞ ⎛ σ11 σ12 σ13 ⎞ ⎛ E1 ⎞

⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ J 2 ⎟ = ⎜ σ21 σ22 σ23 ⎟ ⎜ E2 ⎟ (2.8)
⎜ J 3 ⎟ ⎜ σ31 σ32 σ33 ⎟⎠ ⎜⎝ E3 ⎟⎠
⎝ ⎠ ⎝
or
Ji = ∑σ E .
j
ij j (2.9)
A third-rank tensor is exemplified by piezoelectric coefficients, providing a relation of the induced

strain x with the applied field E:
x = dE. (2.10)
Since E and x are first-rank and second-rank tensors, respectively, d should have a third-rank tensor
form represented as
X jk = ∑d i
ijk Ei . (2.11)
The d tensor is composed of three layers of symmetrical matrices:
⎛ d111 d112 d113 ⎞

⎜ ⎟
1st layer (i = 1) ⎜ d121 d122 d123 ⎟
⎜ d131 d133 ⎟⎠
⎝ d132

Mathematical Treatment of Ferroelectrics 27
⎛ d211 d212 d213 ⎞

⎜ ⎟
2nd layer (i = 2) ⎜ d221 d222 d223 ⎟
⎜ d231 d233 ⎟⎠
⎝ d232
⎛ d311 d312 d313 ⎞

3rd layer (i = 3) ⎜ d321 d322 d323 ⎟ . (2.12)
⎜ ⎟
⎝ d331 d332 d333 ⎠
Generally speaking, if two physical properties are represented using tensors of p-rank and q-rank,
the quantity which combines the two properties in a linear relation is also represented by a tensor
of (p + q)-rank.
2.1.3 CRYSTAL SYMMETRY AND TENSOR FORM

A physical property measured along two different directions must be equal if these two directions
are crystallographically equivalent. This consideration sometimes reduces the number of indepen-
dent tensor components representing the above property.
Let us again take electric conductivity as an example of a second-rank tensor. If the current
density J in an (x, y, z) coordinate system is described in an (x′, y′, z′) system, as J1′, J2′, and J3′ are
related using a unitary transformation matrix as follows:
⎛ J1′ ⎞ ⎛ a11 a12 a13 ⎞ ⎛ J1 ⎞

⎜ J′⎟ = ⎜ a a22 a23 ⎟ ⎜ J 2 ⎟ . (2.13)
⎜ 2 ⎟ ⎜ 21 ⎟⎜ ⎟
⎝ J 3′ ⎠ ⎝ a31 a32 a33 ⎠ ⎝ J 3 ⎠
The electric field is transformed in the same way:
⎛ E1′ ⎞ ⎛ a11 a12 a13 ⎞ ⎛ E1 ⎞

⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ E2′ ⎟ = ⎜ a21 a22 a23 ⎟ ⎜ E2 ⎟ (2.14)
⎜ E3′ ⎟ ⎜ a31 a33 ⎟⎠ ⎜⎝ E3 ⎟⎠
⎝ ⎠ ⎝ a32
or
Ei = ∑a E .
j
ij j (2.15)
Then, we can calculate the corresponding σ′ tensor defined by
⎛ J1′ ⎞ ⎛ E1′ ⎞
⎜ J ′ ⎟ = σ′ ⎜ E ′ ⎟ . (2.16)
⎜ 2⎟ ⎜ 2⎟
⎝ J 3′ ⎠ ⎝ E3′ ⎠
Inserting Equations 2.13 and 2.14 into Equation 2.16,
⎛ a11 a12 a13 ⎞ ⎛ J1 ⎞ ⎛ a11 a12 a13 ⎞ ⎛ E1 ⎞

⎜ ⎟⎜ ⎟ ⎜ ⎟⎜ ⎟
⎜ a21 a22 a23 ⎟ ⎜ J 2 ⎟ = σ′ ⎜ a21 a22 a23 ⎟ ⎜ E2 ⎟
⎜ a31 a33 ⎟⎠ ⎜⎝ J 3 ⎟⎠ ⎜ a31 a33 ⎟⎠ ⎜⎝ E3 ⎟⎠
⎝ a32 ⎝ a32

then
−1
⎛ J1 ⎞ ⎛ a11 a12 a13 ⎞ ⎛ a11 a12 a13 ⎞ ⎛ E1 ⎞
⎜ J ⎟ =⎜ a a22 a23 ⎟ σ′ ⎜ a21 a22 a23 ⎟ ⎜ E2 ⎟ .
⎜ 2 ⎟ ⎜ 21 ⎟ ⎜ ⎟⎜ ⎟
⎝ J 3 ⎠ ⎝ a31 a32 a33 ⎠ ⎝ a31 a32 a33 ⎠ ⎝ E3 ⎠
From the definition of Equation 2.8,

−1
′
⎛ σ11 ′
σ12 ′ ⎞ ⎛ a11
σ13 a12 a13 ⎞ ⎛ σ11 σ12 σ13 ⎞ ⎛ a11 a12 a13 ⎞
⎜ ⎟ ⎜ ⎟⎜ ⎟⎜ ⎟
′
⎜ σ21 ′
σ22 ′
σ23 ⎟ = ⎜ a21 a22 a23 ⎟⎜ σ21 σ22 σ23 ⎟ ⎜ a21 a22 a23 ⎟
⎜ σ′31 σ′′32 σ′33 ⎟⎠ ⎜⎝ a31 a33 ⎟⎜ σ33 ⎟⎠ ⎜⎝ a31 a33 ⎟⎠
⎝ a32 ⎠ ⎝ σ31 σ32 a32
⎛ a11 a12 a13 ⎞ ⎛ σ11 σ12 σ13 ⎞ ⎛ a11 a21 a31 ⎞

⎜ ⎟⎜ ⎟⎜ ⎟
= ⎜ a21 a22 a23 ⎟ ⎜ σ21 σ22 σ23 ⎟ ⎜ a12 a22 a32 ⎟ (2.17)
⎜ a31 a33 ⎟⎠ ⎜⎝ σ31 σ32 σ33 ⎟⎠ ⎜⎝ a13 a33 ⎟⎠
⎝ a32 a23
or
σ′ij = ∑a a σ .
k ,l
ik jl kl (2.18)
When the crystal has a twofold axis along the z-axis, the electric conductivity should have the same
tensor form in terms of the transformation (Equation 2.6 for θ = 180°):
⎛ −1 0 0⎞
⎜ ⎟
⎜0 −1 0⎟
⎜0 1 ⎟⎠
⎝ 0
From the condition
⎛ σ11
′ σ12
′ σ13
′ ⎞ ⎛ −1 0 0⎞ ⎛ σ11 σ12 σ13 ⎞ ⎛ −1 0 0⎞
⎜ σ′ σ′22 σ′23 ⎟ = ⎜ 0 −1 0⎟ ⎜ σ 21 σ 22 σ 23 ⎟ ⎜ 0 −1 0⎟
⎜ 21 ⎟ ⎜ ⎟⎜ ⎟⎜ ⎟
⎝ σ′31 σ′′32 σ′33 ⎠ ⎝ 0 0 1⎠ ⎝ σ31 σ32 σ33 ⎠ ⎝ 0 0 1⎠
⎛ a11 a12 a13 ⎞ ⎛ σ11 σ12 σ13 ⎞ ⎛ a11 a21 a31 ⎞

= ⎜ a21 a22 a23 ⎟ ⎜ σ 21 σ 22 σ 23 ⎟ ⎜ a12 a22 a32 ⎟
⎜ ⎟⎜ ⎟⎜ ⎟
⎝ a31 a32 a33 ⎠ ⎝ σ31 σ32 σ33 ⎠ ⎝ a13 a23 a33 ⎠
⎛ σ11 σ12 −σ13 ⎞

= ⎜ σ 21 σ 22 −σ 23 ⎟ . (2.19)
⎜ ⎟
⎝ −σ31 −σ32 σ33 ⎠
Since this crystal has twofold symmetry, σ′ ≡ σ must be held for this 180° rotation transformation.
Thus, following equivalency should be derived:
⎛ σ11 σ12 −σ13 ⎞ ⎛ σ11 σ12 σ13 ⎞

⎜ σ σ 22 −σ 23 ⎟ = ⎜ σ 21 σ 22 σ 23 ⎟ . (2.20)
⎜ 21 ⎟ ⎜ ⎟
⎝ −σ31 −σ32 σ33 ⎠ ⎝ σ31 σ32 σ33 ⎠

Using the logic that x = −x → x = 0, we can obtain

σ31 = σ13 = σ32 = σ 23 = 0
σ11, σ 22 , σ33 ≠ 0
σ12 = σ 21. (2.21)
It is very important to note that most physical constants have a symmetric tensor form. (The proof
involves thermodynamical considerations that are beyond the scope of this textbook. Refer to Ref. [1].)
For a third-rank tensor such as the piezoelectric tensor, the transformation due to a change in
coordinate system is represented by
′ =
dijk ∑a a
l ,m,n
il a d
jm kn lmn . (2.22)
When the crystal has a fourfold axis along z-axis, for example, the transformation matrix is given
(Equation 2.6 for θ = 90°) by
⎛0 1 0⎞
⎜ −1 0 0⎟ .
⎜ ⎟
⎝0 0 1⎠
Considering the tensor symmetry with m and n such that d123 = d132 and d213 = d231 (each matrix of
the ith layer of the d tensor is symmetrical), we can obtain (Refer to Example Problem 2.1)
d111 = d222 = d112 = d121 = d211 = d221 = d212 = d122
= d331 = d313 = d133 = d332 = d323 = d233
= d312 = d321 = 0
d333 ≠ 0
d311 = d322
d113 = d131 = d223 = d232
d123 = d132 = − d213 = − d231. (2.23)
Then, we get the d tensor as follows:

⎛ 0 0 d131 ⎞
⎜ ⎟
1st layer (i = 1) ⎜ 0 0 d123 ⎟
⎜ 0 ⎠⎟
⎝ d131 d123
⎛ 0 0 −d123 ⎞
⎜ ⎟
2nd layer (i = 2) ⎜ 0 0 d131 ⎟
⎜ 0 ⎠⎟
⎝ −d123 d131
⎛ d311 0 0 ⎞
3rd layer (i = 3) ⎜ 0 d311 0 ⎟. (2.24)
⎜ ⎟
⎝ 0 0 d333 ⎠

EXAMPLE PROBLEM 2.1

For a third-rank piezoelectric tensor, the transformation due to a change in coordinate system is
represented by
′ =
dijk ∑a a
l , m, n
il a d
jm kn lmn .
When a crystal has a fourfold axis along z-axis (point group 4), the transformation matrix is
given by
⎛0 1 0⎞
⎜ −1 0 0⎟ .
⎜ ⎟
⎜⎝ 0 0 1⎟⎠
1. Derive the relationship of Equation 2.23

2. Further, when a crystal has additional mirror symmetry normal to x-axis and y-axis (point
group 4 mm), derive the tensor components. The transformation matrix should be
⎛ −1 0 0⎞ ⎛1 0 0⎞
⎜0 1 0⎟ and ⎜ 0 −1 0⎟ .
⎜ ⎟ ⎜ ⎟
⎜⎝ 0 0 1⎟⎠ ⎜⎝ 0 0 1⎟⎠
Solution
1. Taking into account the tensor symmetry in terms of the second and third suffixes such that
d123 = d132 and d213 = d231 (each matrix of the ith layer of the d tensor is symmetrical), we have
18 independent dijks. Notice initially a12 = 1, a21 = −1, a33 = 1, which means that the second
suffix is automatically determined (2, 1, or 3) when the first suffix is given (1, 2, or 3).
′ =
d111 ∑a a
l ,m ,n
a d
1l 1m 1n 1mn
= a12 a12 a12 d222
= (+1)(+1)(+1)d222 = d222 (P2.1.1)
′ ≡ d111, we obtain d111 = d222.

Since d111
′ =
d122 ∑a a
l ,m ,n
a d
1l 2m 2n 1mn
= a12 a21a21d211
= (+1)( −1)( −1)d211 = d211 (P2.1.2)
′ ≡ d122 , we obtain d122 = d211.

Since d122
′ = a12 a33a33d233
d133
= (+1)(+1)(+1)d233 = d233 (P2.1.3)
′ ≡ d133, we obtain d133 = d233.

Since d133
′ = a12 a21a33d213
d123
= (+1)( −1)(+1)d213 = − d213 (P2.1.4)

′ ≡ d123 , we obtain d123 = −d213 = d132 = −d231.

Since d123
′ = a12 a33a12 d232

d131
= (+1)(+1)(+1)d232 = d232 (P2.1.5)
′ ≡ d131, we obtain d131 = d232 = d113 = d223.

Since d131
′ = a12 a12 a21d221

d112
=(+1)(+1)( − 1)d221 = − d221 (P2.1.6)
′ ≡ d112 , we obtain d112 = −d221 = d121 = −d212.

Since d112
′ = a21a12 a12 d122

d211
= ( −1)(+1)(+1)d122 = − d122 (P2.1.7)
′ ≡ d211, we obtain d211 = −d122.

Since d211
′ = a21a21a21d111
d222
= ( −1)( −1)( −1)d111 = − d111 (P2.1.8)
′ ≡ d222 , we obtain d222 = −d111.

Since d222
′ = a21a33a33d133
d233
= ( −1) (+1)(+1) d133 = − d133 (P2.1.9)
′ ≡ d233 , we obtain d233 = −d133.

Since d233
′ = a21a21a33d113
d223
= ( −1)( −1)(+1) d113 = d113 (P2.1.10)
′ ≡ d223 , we obtain d223 = d113.

Since d223
′ = a21a33a12 d132
d231
= ( −1)(+1)(+1) d132 = − d132 (P2.1.11)
′ ≡ d231, we obtain d231 = −d132.

Since d231
′ = a21a12 a21d121
d212
= ( −1)(+1)( −1) d121 = d121 (P2.1.12)
′ ≡ d212 , we obtain d212 = d121.

Since d212
′ = a33a12 a12 d322

d311
= (+1)(+1)(+1) d322 = d322 (P2.1.13)
′ ≡ d311, we obtain d311 = d322.

Since d311
′ = a33a21a21d311
d322
= (+1) ( −1) ( −1) d311 = d311 (P2.1.14)

′ ≡ d322 , we obtain d322 = d311.

Since d322
′ = a33a33a33d333
d333
= (+1)(+1)(+1) d333 = d333 (P2.1.15)
′ ≡ d333 , we obtain trivial d333 = d333.

Since d333
′ = a33a21a33d313
d323
= (+1)( −1)(+1) d313 = − d313 (P2.1.16)
′ ≡ d323, we obtain d323 = −d313.

Since d323
′ = a33a33a12 d332
d331
= (+1)(+1)(+1) d332 = d332 (P2.1.17)
′ ≡ d331, we obtain d331 = d332.

Since d331
′ = a33a12 a21d321
d312
= (+1) (+1) ( −1)d321 = − d321 (P2.1.18)
′ ≡ d312, we obtain d312 = −d321.

Since d312
From Equations P2.1.2 and P2.1.7, we obtain d122 = d211 = −d122 = 0.
From Equations P2.1.6 and P2.1.12, we obtain d112 = −d221 = d121 = −d212 = d212 = 0.
From Equations P2.1.16 and P2.1.17, we obtain d323 = −d313 = −d331 = d331 = d332 = 0.
From Equations P2.1.18, we obtain d312 = d321 = −d321 = 0.
From Equations P2.1.4, we obtain d123 = −d213 = d132 = −d231.
Now we obtained the results
d111 = d222 = d112 = d121 = d211 = d221 = d212 = d122
= d331 = d313 = d133 = d332 = d323 = d233
= d312 = d321 = 0
d333 ≠ 0
d311 = d322
d113 = d131 = d223 = d232
d123 = d132 = − d213 = − d231.
2. We consider d113 and d123 for the transformation matrix; a11 = −1, a22 = 1, a33 = 1.
′ = a11a11a33d113
d113
= ( −1) ( −1) (+1)d113 = d113 (P2.1.19)
′ ≡ d113, we obtain trivial d113 = d113.

Since d113
′ = a11a22 a33d123
d123
= ( −1) ( +1) ( +1) d123 = − d123 (P2.1.20)

′ ≡ d123, we obtain d123 = −d123 = 0.

Since d123
The mirror symmetry further eliminates
d123 = d132 = − d213 = − d231 = 0.
2.1.4 REDUCTION OF THE TENSOR (MATRIX NOTATION)

A general third-rank tensor has 33 = 27 independent components. Since dijk is symmetrical in j and
k some of the coefficients can be eliminated, leaving 18 independent dijk coefficients; this facilitates
the use of matrix notation.
So far all the equations have been developed in full tensor notation. But when calculating actual
properties, it is advantageous to reduce the number of suffixes as much as possible. This is done by
defining new symbols, for instance, d21 = d211 and d14 = 2d123: The second and third suffixes in the
full tensor notation are replaced by a single suffix 1 to 6 in matrix notation, as follows:
Tensor notation 11 22 33 23, 32 31, 13 12, 21

Matrix notation 1 2 3 4 5 6
In terms of these new symbols the array Equation 2.12 is rewritten as
⎛ ⎛1⎞ ⎛1⎞ ⎞
⎜ d11 ⎜ 2 ⎟ d16 ⎜ 2 ⎟ d15 ⎟
⎜ ⎝ ⎠ ⎝ ⎠ ⎟
⎜⎛ 1 ⎞ ⎛1⎞ ⎟
1st layer (i = 1) ⎜ ⎜ ⎟ d16 d12 ⎜ ⎟ d14 ⎟
⎜⎝ 2 ⎠ ⎝2⎠ ⎟
⎜⎛ 1 ⎞ ⎛1⎞ ⎟
⎜⎜ ⎜ ⎟ d15 ⎜ 2 ⎟ d14 d12 ⎟⎟
⎝⎝ 2 ⎠ ⎝ ⎠ ⎠
⎛ ⎛1⎞ ⎛1⎞ ⎞
⎜ d21 ⎜ 2 ⎟ d26 ⎜ 2 ⎟ d25 ⎟
⎜ ⎝ ⎠ ⎝ ⎠ ⎟
⎜⎛ 1 ⎞ ⎛1⎞ ⎟
2nd layer (i = 2) ⎜ ⎜ ⎟ d26 d22 ⎜ 2 ⎟ d24 ⎟
⎜⎝ 2 ⎠ ⎝ ⎠ ⎟
⎜⎛ 1 ⎞ ⎛1⎞ ⎟
⎜⎜ ⎜ ⎟ d25 ⎜ 2 ⎟ d24 d23 ⎟⎟
⎝⎝ 2 ⎠ ⎝ ⎠ ⎠
⎛ ⎛ 1⎞ ⎛ 1⎞ ⎞
⎜ d31 ⎜⎝ 2 ⎟⎠ d36 ⎜⎝ 2 ⎟⎠ d35 ⎟
⎜ ⎟
⎜⎛ 1⎞ ⎛ 1⎞ ⎟
3rd layer (i = 3) ⎜ ⎜ ⎟ d36 d32 ⎜⎝ 2 ⎟⎠ d34 ⎟ . (2.25)
⎝ 2⎠
⎜ ⎟
⎜⎛ 1⎞ ⎛ 1⎞ ⎟
⎜ ⎜⎝ ⎟⎠ d35 ⎜⎝ 2 ⎟⎠ d34 d33 ⎟
⎝ 2 ⎠
The last two suffixes in the tensor notation correspond to those of the strain components; therefore,
for consistency, we make the following change in the notation for the strain components:

⎛ ⎛ 1⎞ ⎛ 1⎞ ⎞
⎜ x2 ⎜⎝ 2 ⎟⎠ x6 ⎜⎝ 2 ⎟⎠ x5 ⎟
⎛ x11 x12 x31 ⎞ ⎜ ⎟
⎜x ⎜⎛ 1⎞ ⎛ 1⎞ ⎟
x22 x23 ⎟ = ⎜ ⎜ ⎟ x6 x2 ⎜⎝ 2 ⎟⎠ x4 ⎟ . (2.26)
⎜ 12 ⎟ ⎝ 2⎠
⎝ x31 x23 x33 ⎠ ⎜ ⎟
⎜⎛ 1⎞ ⎛ 1⎞ ⎟
⎜ ⎜⎝ ⎟⎠ x5 ⎜⎝ 2 ⎟⎠ x4 x3 ⎟
⎝ 2 ⎠
The reason for the (1/2)s in the substitution (Equation 2.26) is due to the cancellation with (1/2)s in
Equation 2.25. Then, we have
Xj = ∑d E
i
ij i (i = 1,2,3; j = 1,2,...,6) (2.27)
or
⎛ x1 ⎞ ⎛ d11 d21 d31 ⎞
⎜ x ⎟ ⎜d d22 d32 ⎟
⎜ 2 ⎟ ⎜ 12 ⎟ ⎛ E1 ⎞
⎜ x3 ⎟ ⎜ d13 d23 d33 ⎟ ⎜ ⎟
⎜ x ⎟ = ⎜d E2 . (2.28)
d24 d34 ⎟ ⎜ ⎟
⎜ 4 ⎟ ⎜ 14 ⎟ ⎝ E3 ⎠
⎜ x5 ⎟ ⎜ d15 d25 d35 ⎟
⎜ ⎟ ⎜ ⎟
⎝ x6 ⎠ ⎝ d16 d26 d36 ⎠
Concerning the stress components, the (1/2)s are unnecessary:
⎛ X11 X12 X31 ⎞ ⎛ X1 X6 X5 ⎞

⎜ ⎟ ⎜ ⎟
⎜ X12 X 22 X 23 ⎟ = ⎜ X6 X2 X4 ⎟ (2.29)
⎜ X31 X33 ⎟⎠ ⎜⎝ X 5 X3 ⎠⎟
⎝ X 23 X4
The matrix notation has the advantage of compactness over the tensor notation, and it makes it easy
to display the coefficients on a plane diagram (i.e., on a paper. 3-D dijk expression can be described
by a 3-D holographic book in the future.). However, it must be remembered that in spite of their
form, the dij’s do not transform like the components of a second-rank tensor. An example of a piezo-
electric matrix for the point group 4 can be written (from Equation 2.24) as
⎛ 0 0 0 d14 d15 0⎞
⎜ 0 0 0 d15 − d14 0⎟ . (2.30)
⎜ ⎟
⎝ d31 d31 d33 0 0 0⎠
For the point group 4 mm, d14 = −d14 = 0.
2.1.5 MATRIX NOTATION OF ELECTROSTRICTIVE COEFFICIENTS

In the solid state theoretical treatment of the phenomenon of piezoelectricity or electrostriction, the
strain xkl is expressed in terms of the electric field Ei or electric polarization Pi as follows (refer to
Section 2.2):
X kl = ∑d i
ikl Ei + ∑M
i, j
ijkl Ei E j
= ∑g i
ikl i P ∑+ ∑Q i, j
ijkl i PPj , (2.31)

where
dikl and gikl are called the piezoelectric coefficients
Mijkl and Qijkl the electrostrictive coefficients
Since the E and x are first-rank and second-rank tensors, respectively, d should be a third-rank
tensor. However, EiEj and PiPj are not tensors, precisely speaking, but we treat a combination of
(E12, E22, E32, E2E3, E3E1, E1E2) as if these are equivalent to a second-rank tensor. Note that these
six components are not independent at all. This is a smart way to convert the nonlinear behavior
(electrostriction) to a linear algebra problem.
Using a similar reduction of the notation for the electrostrictive coefficients Mijkl, we get the fol-
lowing equation corresponding to Equation 2.31:
⎛ x1 ⎞ ⎛ d11 d21 d31 ⎞

⎜ x ⎟ ⎜d d22 d32 ⎟
⎜ 2 ⎟ ⎜ 12 ⎟ ⎛ E1 ⎞
⎜ x3 ⎟ ⎜ d13 d23 d33 ⎟ ⎜ ⎟
⎜ x ⎟ = ⎜d d24
E2
d34 ⎟ ⎜ ⎟
⎜ 4 ⎟ ⎜ 14 ⎟ ⎝ E3 ⎠
⎜ x5 ⎟ ⎜ d15 d25 d35 ⎟
⎜ ⎟ ⎜ ⎟
⎝ x6 ⎠ ⎝ d16 d26 d36 ⎠
⎛ M11 M 21 M31 M 41 M 51 M61 ⎞ ⎛ E12 ⎞

⎜M M 22 M32 M 42 M 52 M62 ⎟ ⎜ E22 ⎟
⎜ 12 ⎟⎜ ⎟
⎜ M13 M 23 M33 M 43 M 53 M63 ⎟ ⎜ E32 ⎟
+⎜ ⎜ ⎟.
M64 ⎟ ⎜ E2 E3 ⎟
(2.32)
M M 24 M34 M 44 M 54
⎜ 14 ⎟
⎜ M15 M 25 M35 M 45 M 55 M65 ⎟ ⎜ E3 E1 ⎟
⎜ ⎟⎜ ⎟
⎝ M16 M 26 M36 M 46 M 56 M66 ⎠ ⎝ E1E2 ⎠
Tables 2.1 and 2.2 summarize the matrices d and M for all crystallographic point groups.1
EXAMPLE PROBLEM 2.2

Suppose that a shear stress is applied to a square crystal and the crystal is deformed as illustrated
in Figure 2.1. Calculate the induced strain x5 (=2x31).
Solution
Since x5 = 2x31 = tan θ = θ and 1° = π /180 rad, x5 = 0.017. Typical strain in a piezoelectric ceramic
is around 0.1% or 0.001, leading to the shear angle 0.1°.
Note that shear deformation is not accompanied with a volume change, different from the
longitudinal deformation. In order to further understand the shear deformation, try Example
Problem 2.3.
1°
3
F
1
FIGURE 2.1 Shear stress and strain configuration.

TABLE 2.1 Piezoelectric Coefficient Matrix*

dmn = dijk (n = 1, 2, 3)
* dmn = 2dijk (n = 4, 5, 6)
Symbol meanings
Zero component
Nonzero component
Equal component
Equal with opposite signs
–2 times of the connected point
I. Centrosymmetric point group
Point group 1, 2/m, mmm, 4/m, 4/mmm, m3, m3m, 3, 3m, 6/m, 6/mmm.
All components are zero.
II. Noncentrosymmetric point group
Triclinic
Point group 1
(18)
Point group 2 Monoclinic Point group 2

2 || x3
2 || x 2
(Standard)
orientation (8) (8)
Point group m Point group m
m x3
m x2
(Standard)
orientation (10) (10)
Orthorhombic
Point group 222 Point group mm2
(3) (5)
Tetragonal
Point group 4 Point group 4
(4) (4)
Point group 422 Point group 4mm
(1) (3)
Point group 42m
2 || x1
(2)

TABLE 2.1 (continued) Piezoelectric Coefficient Matrix*
Cubic
Point group 432 Point group 43m, 23
(0) (1)
All components are zero
Rhombohedral
(6) (2)
Point group 3m Point group 3m
m x2
m x1
(Standard)
orientation (4) (4)
Hexagonal
Point group 6 Point group 6mm
(4) (3)
(1) (2)
Point group 6m2 Point group 6m2
m x1
m x2
(Standard)
orientation (1) (1)
EXAMPLE PROBLEM 2.3

For a cube-shaped specimen, tensile stress X and compressive stress −X are applied simultane-
–
ously along the (1 0 1) and (1 0 1) axes, respectively (Figure 2.2). When we take the prime coor-
dinates (1′ and 3′) as illustrated in Figure 2.2, the stress tensor is represented as
3¢ 3
1¢ –X
2, 2¢
1
FIGURE 2.2 Application of a pair of stresses X and −X to a cube of material.

TABLE 2.2 Electrostrictive Coefficient Matrix*

Qmn = Qijkl (m, n = 1, 2, 3)
* Qmn = 2Qijkl (m or n = 4, 5, 6)
Qmn = 4Qijkl (m, n = 4, 5, 6)
Symbol meanings
Zero component
Nonzero component
Equal components
Equal with opposite signs

2 times of the connected component
–2 times of the connected component

2(Q11 – Q12)
Triclinic
Point group 1, 1
(36)
Twofold axis || x2 Monoclinic
Standard Point group 2, m, 2/m
orientation Twofold axis || x3
(20) (20)
Orthorhombic
Point group 222, mm2, mmm
(12)
Tetragonal
Point group 4, 4, 4/m Point group 4mm, 42m, 422, 4/mmm
(10) (7)

TABLE 2.2 (continued) Electrostrictive Coefficient Matrix*
Rhombohedral
Point group 3, 3 Point group 3m, 32, 3m
(12) (8)
Point group 6, 6, 6/m Hexagonal Point group 6m2, 6mm, 622, 6/mmm
(8) (6)
Cubic Point group 43m, 432, m3m

Point group 23, m3
(4) (3)
Isotropic
(2)
⎛X 0 0⎞
⎜0 0 0⎟
⎜ ⎟
⎜⎝ 0 0 − X ⎟⎠
Using the transformation matrix A (i.e., θ = −45° rotation along 2′ axis in Figure 2.2)
⎛ cos θ 0 sin θ ⎞
⎜ 0 1 0 ⎟ , calculate AXA−1, and verify that the above stress is equivalent to a pure
⎜ ⎟
⎜⎝ − sin θ 0 cos θ⎟⎠
shear stress in the original (nonprime) coordinates.
Solution
Using θ = −45°, we can obtain the transformed stress representation:

⎛ 1 1 ⎞ ⎛ 1 1 ⎞
0 − 0 −
⎜ 2 2⎟⎛X 0 0⎞ ⎜ 2 2⎟
⎜ ⎟ ⎜ ⎟
−1
AXA = ⎜ 0 1 0 ⎟⎜ 0 0 0⎟ ⎜ 0 1 0 ⎟
⎜ ⎟
⎜ 1 1 ⎟ ⎜⎝ 0 0 − X ⎟⎠ ⎜ 1 1 ⎟
⎜ 0 ⎟ ⎜− 0 ⎟
⎝ 2 2 ⎠ ⎝ 2 2 ⎠
⎛0 0 X⎞
=⎜ 0 0 0⎟ . (P2.3.1)
⎜ ⎟
⎜⎝ X 0 0⎟⎠
The off-diagonal components X13 and X31 have the same magnitude X, and represent a pure
shear stress. Note that a shear stress is equivalent to a combination of extension and contraction
stresses. Only an extensional stress applied along a diagonal direction 1′ may exhibit an appar-
ently similar diagonal distortion of the crystal. However, precisely speaking, without the external
contraction along the 3′ direction, this is not exactly equivalent to the pure shear deformation,
with a volume expansion. The contraction occurs in this case only from the Poisson’s ratio of the
extension.
EXAMPLE PROBLEM 2.4

Barium titanate (BT) shows tetragonal crystal symmetry (point group 4mm) at room tempera-
ture. Therefore, its piezoelectric constant matrix is given by (Equation 2.30):
⎛ 0 0 0 0 d15 0⎞
⎜ 0 0 0 d15 0 0⎟
⎜ ⎟
⎜⎝ d d31 d31 0 0 0⎟⎠
31
a. Calculate the induced strain under an electric field applied along the crystal c-axis (E3).
b. Calculate the induced strain under an electric field applied along the crystal a-axis (E1).
Solution
⎛ x1 ⎞ ⎛ 0 0 d31 ⎞
⎜x ⎟ ⎜ 0 0 d31 ⎟
⎜ 2⎟ ⎜ ⎟ ⎛ E1 ⎞
⎜ x3 ⎟ ⎜ 0 0 d33 ⎟ ⎜ ⎟
⎜x ⎟ = ⎜ 0 d15 0⎟⎟ ⎜⎜ ⎟⎟
E2 (P2.4.1)
⎜ 4⎟ ⎜ ⎝ E3 ⎠
⎜ x5 ⎟ ⎜ d15 0 0⎟
⎜ ⎟ ⎜ ⎟
⎝ x6 ⎠ ⎝ 0 0 0⎠
can be transformed into
x1 = x2 = d31E3 , x3 = d33 E3
x4 = d15 E2 , x5 = d15 E1, x6 = 0. (P2.4.2)
a. When E3 is applied, elongation in the c direction (x3 = d33E3, d33 > 0) and contraction in
the a and b directions (x1 = x2 = d31E3, d31 < 0) are induced. The ratio |d31/d33| corresponds to
Poisson’s ratio σ.
b. When E1 is applied, shear strain x5 (=2x31) = d15E1 is induced. Figure 2.3a illustrates a case
of d15 > 0 and x5 > 0. The deformation can be intuitively understood by the polarization
cant under the electric field.

Polarization
3
2
(a) (b)
1
FIGURE 2.3 (a) Piezoelectric shear strain in the point group tetragonal 4mm, and (b) [111] axis electrostric-
tive strain in the point group cubic m3m.
EXAMPLE PROBLEM 2.5

Lead magnesium niobate (Pb(Mg1/3Nb2/3)O3) exhibits cubic crystal symmetry (point group
m3m) at room temperature and does not show piezoelectricity. However, large electrostriction
is induced under an applied electric field. The relation between the strain and the electric field
is given by
⎛ x1 ⎞ ⎛ M11 M 21 M31 M 41 M 51 M61 ⎞ ⎛ E12 ⎞

⎜x ⎟ ⎜M M 22 M32 M 42 M 52 M62 ⎟ ⎜ E22 ⎟
⎜ 2 ⎟ ⎜ 12 ⎟⎜ ⎟
⎜ x3 ⎟ ⎜ M13 M 23 M33 M 43 M 53 M63 ⎟ ⎜ E32 ⎟
⎜x ⎟ = ⎜M M 24 M34 M 44 M 54
⎜ ⎟
M64 ⎟⎟ ⎜ E2 E3 ⎟
(P2.5.1)
⎜ 4 ⎟ ⎜ 14
⎜ x5 ⎟ ⎜ M15 M 25 M35 M 45 M 55 M65 ⎟ ⎜ E3 E1 ⎟
⎜ ⎟ ⎜ ⎟⎜ ⎟
⎝ x6 ⎠ ⎝ M16 M 26 M36 M 46 M 56 M66 ⎠ ⎝ E1E2 ⎠
in a matrix representation. Calculate the induced strain under an electric field applied along the
[111] direction (based on a cubic perovskite coordinate).
Solution
The electric field along the [111] direction, E[111], is represented as ( E[111] / 3, E[111] / 3, E[111] / 3) .
Substituting E1 = E2 = E3 = E[111] / 3 into Equation P2.5.1, we obtain
x1 = x2 = x3 = (M11 + 2M12 )E[111]

2
/3 (= x11 = x22 = x33 )
x4 = x5 = x6 = M 44 E[111]
2
/3 (= 2 x23 = 2 x31 = 2 x12 ). (P2.5.2)
The distortion is illustrated in Figure 2.3b. The strain x induced along an arbitrary direction is
given by
x = ∑ xij lil j (P2.5.3)

i, j
where li is a direction cosine with respect to the i-axis. Therefore, the strain induced along the
[111] direction, x[111]//, is given by
X[111]/ / = ∑ Xij (1/ 3)(1/ 3 )

i, j
= ⎡⎣ x1 + x2 + x3 + 2 (x4 /2 + x5 /2 + x6 /2 )⎤⎦ /3
= (M11 + 2M12 + M 44 )E[111]

2
/3. (P2.5.4)

On the other hand, the strain induced perpendicular to the [111] direction, x[111]⊥, is calculated in
a similar fashion as
x[111]⊥ = (M11 + 2M12 – M 44 /2 ) E[111]

2
/3. (P2.5.5)
Figure 2.3b shows the distortion schematically. It is important to note that the volumetric strain
(ΔV/V) given by
x⎡⎣111⎤⎦ / / + 2x[111]⊥ = (M11 + 2 M12 )E[111]

2
. (P2.5.6)
Note that this volumetric strain is the same for (ΔV/V) under an electric field E1:
x1 + 2x2 = (M11 + 2 M12 )E12 , (P2.5.7)
leading to the conclusion; (ΔV/V) is independent of the applied field direction.
2.2 PHENOMENOLOGY OF FERROELECTRICITY

2.2.1 FUNDAMENTALS OF PHENOMENOLOGY
A thermodynamic phenomenological theory is discussed basically in the form of expansion series
of the free energy as a function of the physical properties; one of polarization P and electric field E,
one of temperature T and entropy S, one of stress X and strain x, and if applicable, one of magnetic
field H and magnetization M. In our ferroelectric discussion, the last parameters will be neglected.
This derivation process is the most frequently asked question from the readers.
Polarization expansion
The free energy can be expanded in general as follows:
F (P ) = a1P + a2 P 2 + a3 P 3 + a4 P 4 + a5 P 5 + a6 P 6 +
We assume that the free energy of the crystal should not change with polarization reversal (P → −P).
Otherwise, the charge or permittivity in the capacitance would be changed according to the capaci-
tor orientation/upside down, which may cause serious practical problems in electronic equipment.
From F(P) = F(−P), the series should not contain terms in odd powers of P. Thus, the expansion
series includes only even powers of P:
F (P ) = a2P 2 + a4 P 4 + a6 P 6 +
Temperature expansion
Next, we take into account the expansion series in terms of P and temperature:
F (P, T ) = a2 P 2 + a4 P 4 + a6 P 6 + + b1T + b2T 2 + … + c1TP 2 +
From S = − (∂F ∂T ) = −b1, and that a constant entropy is meaningless, we take b1 = 0. The term
b2T 2 is a higher order term to be neglected. Thus we adopt only c1TP2. Note that a possible term TP
is omitted from the reason F(P) = F(−P) again. It is important to understand that the product TP2 of
the two parameters (P2 and T) explain the coupling effect; that is, T change causes P change to keep
the same free energy (this effect is called “pyroelectric” effect), or E application causes T change
(this is called “electrocaloric” effect). By combining a2P2 and c1TP2, we introduce
(1/2)α = a2 P 2 + c1TP 2 = (1/2)⎛⎜⎝ Tε−CT0 ⎞⎟⎠ .

0

We also introduce the following notations:
(1/4)β = a4
(1/6)γ = a6 .
Stress expansion
Now we introduce the stress expansion series:
F (P, T , X ) = (1/2 )α (T )P 2 + (1/4 ) βP 4 + (1/6 ) γP 6 +
+ d1 X + d2 X 2 + + e1P 2 X + [α(T ) = (T − T0 )/ ε 0C ]
From x = − (∂F ∂X ) = −d1, and that constant strain is meaningless, we take d1 = 0 (strain origin).
P2 X is the fundamental electromechanical coupling (i.e., electrostrictive coupling), which explains
the polarization generation under stress, or strain generation under electric field. This argument is
valid only when the spontaneous polarization exists in piezoelectric phase, and the polarization
disappears in paraelectric phase. It is not valid in quartz, where the piezoelectric phase does not
have the spontaneous polarization (actually, quartz is not a ferroelectric). Odd power of stress X can
remain in the free energy, because the crystal orientation/upside down does not change the sign
of X (X is a tensor, not a vector; negative X means compressive stress). Introducing new notations
d2 = −(1/2)s (elastic stiffness), and e1 = −Q (electrostrictive coefficient), we finally obtain
G1 = (1/2 ) ⎡⎣(T − T0 )/ε0C ⎤⎦ P 2 + (1/4 )βP 4 + (1/6 ) γP 6
− (1/2 )sX 2 − QP 2 X .
The above free energy, G1, is particularly called elastic Gibbs energy.
2.2.2 LANDAU THEORY OF THE PHASE TRANSITION

We assume that the Landau free energy F in one dimension is represented in terms of polarization
P (excluding stress terms) as
F (P, T ) = (1/2) α P 2 + (1/4) βP 4 + (1/6 )γP 6 . (2.33)
The coefficients α, β, and γ depend, in general, on the temperature, but as discussed in the previous
section, only α is assumed to be temperature dependent in the following calculation. The phenom-
enological formulation should be applied for the whole temperature range over which the material
is in the paraelectric and ferroelectric states.
Because the spontaneous polarization should be zero in the paraelectric state, the free energy
should be zero in the paraelectric phase at any temperature above its Curie temperature (or the phase
transition temperature). To stabilize the ferroelectric state, the free energy for a certain polariza-
tion P should be lower than “zero.” Otherwise, the paraelectric state should be realized. Thus, at
least, the coefficient α of the P2 term must be negative for the polarized state to be stable; while in
the paraelectric state it must be positive passing through zero at some temperature T0 (Curie–Weiss
temperature). From this concept, we get a linear relation
α = (T – T0 )/ε 0C , (2.34)

where
C is taken as a positive constant called the Curie–Weiss constant
T0 is equal to or lower than the actual transition temperature TC (Curie temperature)
ε0 is the vacuum permittivity
The temperature dependence of α is related on a microscopic level to the temperature dependence

of the ionic polarizability coupled with thermal expansion and other effects of anharmonic lattice
interactions. Refer to the discussion in Section 1.2.
The equilibrium polarization in an electric field E should satisfy the condition:
(∂F/∂P ) = E = αP + βP 3 + γP 5 . (2.35)
With no electric field applied, Equation 2.35 provides two cases:
P(α + βP 2 + γP 4 ) = 0, (2.36)
i. P = 0 → This trivial solution corresponds to a paraelectric state.

ii. α + βP2 + γP4 = 0 → This finite polarization solution corresponds to a ferroelectric state.
Second-order transition
When β is positive, γ is often neglected because nothing special is added by this term. There are not
many examples which show this “second-order” transition, but this description provides intuitive
ideas on the phase transition because of its mathematical simplicity. Triglycine sulphate (TGS) is an
example of a ferroelectric exhibiting the second-order transition.
The polarization for zero applied field is obtained from Equation 2.35 as
α PS + βPS3 = 0 [α = (T – T0 )/ε 0C ], (2.37)
so that either PS = 0 or PS2 = −α /β = (T0 − T )/β ε 0C .

For T > T0, the unique solution PS = 0 is obtained. For T < T0 the minimum of the Landau free
energy is obtained at
T0 − T
PS = ± . (2.38)
βε0C
The phase transition occurs at TC = T0 and the polarization goes continuously to zero at this tempera-
ture; this is called a second-order transition.
The relative permittivity ε is calculated as
1/ε = ε 0 /(∂P / ∂E ) = ε 0 (α + 3β P 2 ). (2.39)
Then, for T > T0, P = 0:
ε = 1/ε 0α = C /(T − T0 ) (T > T0 ). (2.40)
For T < T0, PS2 = −α/β = (T0 − T)/βε0 C:
1/ε = ε 0 (α + 3βP 2 ) = ε 0 [α + 3β( −α /β)] = −2ε 0α .
ε = −1/2ε 0α = C / ⎡⎣2 (T0 − T )⎤⎦ (T < T0 ). (2.41)

Permittivity ε Permittivity ε
Physical properties
Physical properties
Spontaneous Inverse Spontaneous
PS permittivity PS Inverse
1/ε permittivity
1/ε
T1
Tc Temperature To Tc Temperature
(a) (Curie temp.) (b) (Curie temp.)
FIGURE 2.4 Phase transitions in a ferroelectric: (a) second-order and (b) first-order.
Figure 2.4a shows the variations of PS and ε with temperature. It is notable that the permittivity
becomes infinite at the transition temperature, and that the slope of the inverse permittivity in the
ferroelectric phase is 2 times of that in the paraelectric phase.
First-order transition
When β is negative in Equation 2.33 and γ is taken positive, the transition becomes first order. The
equilibrium condition for E = 0 in Equation 2.42 leads to either PS = 0 or Equation 2.43:
[(T − T0 )/ ε 0C ]PS + βPS3 + γPS5 = 0 (2.42)
PS2 = [ −β + β2 − 4 γ (T − T0 )/ε 0C ]/2 γ. (2.43)
Frequently Asked Question
Is PS2 = [ −β − β 2 − 4γ (T − T0 )/ε 0C ]/2γ not another root for solving Equation 2.42?
Answer
Since (T − T0) < 0 and γ > 0, β 2 − 4γ (T − T0 )/ε 0C > β 2 = −β (or β ). Note that β < 0. Thus,
[ −β − β 2 − 4γ (T − T0 )/ε 0C ]/2γ < 0 , which is contradictory with Ps2. The spontaneous polarization
should be positive or negative, but still be a real number, never be an imaginary number.
The transition temperature TC is obtained from the condition that the free energies of the parael-
ectric and ferroelectric phases are equal: i.e., F = 0:
(1/2)[(T − T0 )/ ε0C ] PS2 + (1/4)βPS4 + (1/6 )γPS6 = 0. (2.44)
Therefore, we can derive the Curie temperature (Refer to Example Problem 2.6):
TC = T0 + (3/16 )(β2 ε 0C / γ ). (2.45)
Note that the Curie temperature TC is slightly higher than the Curie–Weiss temperature T0, and that
a discrete jump of PS appears at TC. Also, the permittivity exhibits a finite maximum at TC for a
first-order transition (Figure 2.4b). BT is an example of a ferroelectric that undergoes a first-order
phase transition.
Figure 2.4b also shows the variation of ε with temperature. The slope of the inverse permittivity
in the ferroelectric phase is 8 times of that in the paraelectric phase (refer to Chapter Problem 2.2).
The extrapolated temperature of the inverse permittivity in the paraelectric phase provides the
Curie–Weiss temperature T0, while the extrapolated temperature of the inverse permittivity in the
ferroelectric phase gives the temperature T1, which are discussed below.

Free energy Free energy

T > TC T > T1
TC < T < T1
T = TC
T = TC
T < TC
T0 < T < TC
P
P T < T0
(a) (b)
FIGURE 2.5 Free energy curves plotted for the second- (a) and first-order (b) phase transitions at various
temperatures.
The free energy curves are plotted for the second- and first-order phase transitions at various
temperatures in Figure 2.5. In the case of β > 0, the phase transition is not associated with a latent
heat, but with a jump of the specific heat; thus, this is called a second-order transition. On the other
hand, in the case of β < 0, the transition exhibits a latent heat, and is called a first-order transi-
tion, where the permittivity shows a maximum and a discontinuity of the spontaneous polarization
appears at TC. When T > T1, there is only one free energy minimum; when TC < T < T1, there are
three potential minima. P = 0 is the lowest energy solution for TC < T < T1, while P = ±PS exhibit
the lowest potential for T0 < T < TC; when T < T0, there are double minima of the free energy, which
correspond to stable spontaneous polarizations. Due to the potential gap between three potential
minima, we observe significant temperature hysteresis of the phase transition between the cycles of
rising and falling temperature.
EXAMPLE PROBLEM 2.6

Verify the difference between the Curie TC and Curie–Weiss T0 temperatures as expressed by
TC = T0 + (3/16)(β2 ε 0C /γ ),
for a first-order phase transition, where the Landau free energy is expanded as
F (P, T ) = (1/2 ) αP 2 + (1/4 )βP 4 + (1/6 )γP 6 ,
α = (T – T0 )/ε 0C.
Hint
The potential minima are obtained from
(∂F/∂P ) = E = αP + βP 3 + γP 5 = 0. (P2.6.1)
This equation is valid for any temperature below and above Curie temperature. There are
generally three minima including P = 0 (F = 0). At the Curie temperature, the free energy at
the nonzero polarization must be equal to the free energy of the paraelectric state; that is, zero
(F = 0). Thus we obtain another condition:
F = (1/2 ) αP 2 + (1/4 )βP 4 + (1/6 )γP 6 = 0. (P2.6.2)
This equation is only valid at the phase transition temperature TC.

Solution
Equations P2.6.1 and P2.6.2 are reduced for nonzero polarizations to
α + βP 2 + γ P 4 = 0, (P2.6.3)
α + (1/2 )βP 2 + (1/3)γ P 4 = 0. (P2.6.4)
Note that Equation P2.6.3 is valid for all temperatures below TC, but Equation P2.6.4 is only valid
at T = TC. Eliminating the P terms from these two equations [3(P2.6.4) – (P2.6.3)]
⎣⎡(3/2 )– 1⎦⎤ β P + ⎣⎡3 – 1⎦⎤ α = 0 → P = −4 α /β.

2 2
Then we obtain the following equation from Equation P2.6.3:
α + β( −4α /β) + γ ( −4α /β)2 = 0. (P2.6.5)
or
−3α + γ × 16α 2 / β2 = 0.
Taking account of α = (T − T0)/ε0 C = (TC − T0)/ε0 C, the Curie temperature is calculated as
TC = T0 + (3/16 )(β2 ε 0C /γ ). (P2.6.6)
2.2.3 PHENOMENOLOGY OF ELECTROSTRICTION

In a ferroelectric whose prototype phase (high-temperature paraelectric phase) is centrosymmetric
and nonpiezoelectric, the piezoelectric coupling term PX is omitted and only the electrostrictive
coupling term P2 X is introduced. The theories for electrostriction in ferroelectrics were formu-
lated in the 1950s by Devonshire2 and Kay.3 Let us assume that the elastic Gibbs energy should be
expanded in a one-dimensional form:
G1 (P, X , T ) = (1/2)αP2 + (1/4 )βP 4 + (1/6)γP 6
− (1/2 )sX 2 − QP 2 X , (α = (T – T0 )/ε 0C ) , (2.46)
where
P, X, T are the polarization, stress, and temperature, respectively
s and Q are called the elastic compliance and the electrostrictive coefficient, respectively
This leads to Equations 2.47 and 2.48:
E = (∂G1 /∂P ) = αP + βP 3 + γP 5 − 2QPX (2.47)
x = −(∂G1 /∂X ) = sX + QP 2. (2.48)
Case I: X = 0
When the external stress is zero, the following equations are derived
E = αP + βP 3 + γP 5 (2.49)
x = QP 2 (2.50)

1/ε 0 ε = α + 3βP 2 + 5γP 4 . (2.51)
If the external electric field is equal to zero (E = 0), two different states are derived P = 0 and
P 2 = (−β + β2 − 4αγ )/2 γ .
i. Paraelectric phase: PS = 0 or P = ε0εE (under small E)
Permittivity: ε = C /(T − T0 )(Curie −Weiss law) (2.52)
Electrostriction: x = Qε 20 ε 2 E 2 . (2.53)
The previously mentioned electrostrictive coefficient M in Equation 2.32 is related to the elec-
trostrictive Q coefficient through
M = Qε 20 ε 2 . (2.54)
ii. Ferroelectric phase: PS2 = (−β + β2 − 4αγ )2 γ or P = PS + ε0εE (under small E)
x = Q( PS + ε 0 εE )2 = QPS2 + 2ε 0 εQPSE + Qε 02 ε 2 E 2 , (2.55)
where we define the spontaneous strain xS and the piezoelectric constant d as
Spontaneous strain: xS = QPS2 (2.56)
Piezoelectric constant: d = 2ε 0 εQPS . (2.57)
We see from Equation 2.57 that piezoelectricity is equivalent to the electrostrictive phenomenon
biased by the spontaneous polarization. The temperature dependences of the spontaneous strain
and the piezoelectric constant are plotted in Figure 2.6.
Case II: X ≠ 0
When a hydrostatic pressure p (X = −p) is applied, the inverse permittivity is changed in proportion
to p:
1/ε 0 ε = α + 3βPS2 + 5γPS4 + 2Qp (Ferroelectric state)
α + 2Qp = (T – T0 + 2Qε 0Cp)/(ε 0C ) (Paraelectric state). (2.58)
Piezoelectric
Physical properties
Spontaneous constant d
strain xs
TC (Curie temp.)
temperature
FIGURE 2.6 Temperature dependence of the spontaneous strain and the piezoelectric constant.

Therefore, the pressure dependence of the Curie–Weiss temperature T0 or the transition temperature
TC is derived as follows:
(∂T0 /∂p) = (∂TC /∂p) = −2Qε 0C . (2.59)
In general, the ferroelectric Curie temperature is decreased with increasing hydrostatic pressure
(i.e., Qh > 0). More precisely, there is a 50°C temperature decrease per 1 GPa hydrostatic pressure
increase in perovskite ferroelectrics.
EXAMPLE PROBLEM 2.7

BT has d33 = 320 × 10 −12 C/N, εc (=ε3) = 800 and Q33 = 0.11 m4 C−2 at room temperature. Estimate
the spontaneous polarization PS.
Solution
Let us use the relationship
d33 = 2ε 0 ε 3Q33 PS . (P2.7.1)
PS = d33 /2 ε 0 ε 3Q33
{
= 320 × 10 −12 ⎡⎣C/N ⎤⎦ / 2 × 8.854 × 10 −12 ⎡⎣ F/m ⎤⎦ × 800 × 0.11 ⎣⎡m 4 C −2 ⎦⎤ }
= 0.21 ⎡⎣C/m 2 ⎤⎦ (P2.7.2)
EXAMPLE PROBLEM 2.8

In the case of a second-order phase transition, the elastic Gibbs energy is expanded in a one-
dimensional form as follows:
G1 (P, X , T ) = (1/2 )αP 2 + (1/4 )βP 4
− (1/2 )sX 2 − QP 2 X , (P2.8.1)
where only the coefficient α is dependent on temperature, α = (T − T0)/ε0 C. Obtain the dielectric
constant, spontaneous polarization, spontaneous strain, and piezoelectric constant as a function
of temperature.
Solution
E = (∂G1 /∂P ) = αP + βP 3 − 2QPX (P2.8.2)
x = −(∂G1 /∂X ) = sX + QP 2 (P2.8.3)
When an external stress is zero, we can deduce the three characteristic equations:
E = αP + βP 3 (P2.8.4)
x = QP 2 (P2.8.5)
1/ε 0 ε = (∂E/∂P ) = α + 3βP 2 . (P2.8.6)

By setting E = 0 initially, we obtain the following two stable states: PS2 = 0 or −α/β.
i. Paraelectric phase −T > T0 : PS = 0
1/ε 0 ε = α, then ε = C ( T − T0 ). (Curie − Weiss law) (P2.8.7)
ii. Ferroelectric phase
−T < T0 : PS = ± (T0 − T ) ε 0Cβ (P2.8.8)
1/ε 0 ε = α + 3βP 2 = −2α, then ε = C /2 (T0 − T ) (P2.8.9)
xS = QPS2 = Q (T0 − T )/ ε 0Cβ. (P2.8.10)
From Equations P2.8.8 and P2.8.9, the piezoelectric constant is obtained as
d = 2ε 0 εQPS
= Q ε 0C β (T0 − T )
−1/ 2
. (P2.8.11)
Note here that both the permittivity and spontaneous polarization are temperature dependent,
and only the electrostrictive coefficient Q is temperature independent. By the way, the elec-
trostrictive coefficient M in Equation 2.54 is also strongly dependent on temperature.
2.2.4 CONVERSE EFFECTS OF ELECTROSTRICTION

So far we have discussed the electric field induced strains, i.e., piezoelectric strain (converse piezo-
electric effect, x = dE) and electrostriction (electrostrictive effect, x = ME2). Let us consider here the
converse effect, that is, the material’s response to an external stress, which is applicable to sensors.
The direct piezoelectric effect is the increase of the spontaneous polarization by an external stress,
and expressed as
ΔP = dX . (2.60)
On the contrary, since an electrostrictive material does not have a spontaneous polarization, it does
not generate any charge under stress, but does exhibit a change in permittivity (see Equation 2.58):
Δ(1/ε 0 ε) = 2QX . (2.61)
This is the converse electrostrictive effect. The converse electrostrictive effect, the stress depen-
dence of the permittivity, is used in stress sensors.4 A bimorph structure which subtracts the static
capacitances of two dielectric ceramic plates can provide superior stress sensitivity and temperature
stability. The capacitance changes of the top and bottom plates have opposite signs for uniaxial
stress and same sign for temperature change. The response speed is limited by the capacitance
measuring frequency to about 1 kHz. Unlike piezoelectric sensors, electrostrictive sensors are effec-
tive in the low frequency range, especially DC. The sensor applications of piezoelectrics will be
discussed in Section 7.2.
2.2.5 TEMPERATURE DEPENDENCE OF ELECTROSTRICTION

We have treated the electrostrictive coefficient Q as a temperature-independent constant in the last
section. How is the actual situation? Several expressions for the electrostrictive coefficient Q have
been given so far. From the data obtained by independent experimental methods such as

3 ×10–2
Q
33
1×10–2
2
Q33 (m4/C2)
–Q
–Q13 (m4/C2)
13
0.5
1
Curie point
–150 –100 50 0 50 100

Temperature (°C)
FIGURE 2.7 Temperature dependence of the electrostrictive constants Q33 and Q31 measured in a single
crystal Pb(Mg1/3Nb2/3)O3.
1. Electric field-induced strain in the paraelectric phase

2. Spontaneous polarization and spontaneous strain (x-ray diffraction) in the ferroelectric
phase
3. d constants from the field-induced strain in the ferroelectric phase or from piezoelectric
resonance
4. Pressure dependence of permittivity in the paraelectric phase, nearly equal values of Q
were obtained. Figure 2.7 shows the temperature dependence of the electrostrictive coeffi-
cients Q33 and Q31 observed for a complex perovskite Pb(Mg1/3Nb2/3)O3 single crystal sam-
ple, whose Curie temperature is near 0°C.5 It is seen that there is no significant anomaly in
the electrostrictive coefficient Q through the temperature range in which the paraelectric to
ferroelectric phase transition occurs and piezoelectricity appears. Q is verified to be almost
temperature independent.
2.3 PHENOMENOLOGY OF ANTIFERROELECTRICITY

2.3.1 ANTIFERROELECTRICS
The previous sections dealt with the case in which the directions of the spontaneous dipoles are
parallel to each other in a crystal (polar crystal). There are cases in which antiparallel orientation
lowers the dipole–dipole interaction energy. Such crystals are called antipolar crystals. Figure 2.8
shows the orientation of the spontaneous electric dipoles in an antipolar state in comparison with a
nonpolar and a polar state. In an antipolar crystal, where the free energy of an antipolar state does
not differ appreciatively from that of a polar state, the application of an external electric field or
mechanical stress may cause a transition of the dipole orientation to a parallel state. Such crystals
are called antiferroelectrics.
Stripe type Checker board type

(a) (b) (c)
FIGURE 2.8 Schematic arrangement of the spontaneous dipoles in (a) nonpolar, (b) polar, and (c) antipolar
materials.

Polarization Polarization
Electric field
Electric field Ec
Polarization
(a) Paraelectric
(b) Ferroelectric Electric field

Et
(c) Antiferroelectric
FIGURE 2.9 Polarization vs. electric field hysteresis curves in (a) paraelectric, (b) ferroelectric, and (c) anti-
ferroelectric materials.
Figure 2.9 shows the relationship between E (applied electric field) and P (induced polarization)
in paraelectric, ferroelectric, and antiferroelectric phases. In a paraelectric phase the P–E relation is
linear; in a ferroelectric phase there appears a hysteresis caused by the transition of the spontaneous
polarization between the positive and negative directions; in an antiferroelectric phase, at low electric
field, the induced polarization is proportional to E, and when E exceeds a certain value Ecrit, the crystal
becomes ferroelectric (electric field-induced phase transition), and the polarization shows hysteresis
with respect to E. After removal of the electric field, the crystal returns to its antipolar state, and hence,
no spontaneous polarization can be observed as a whole. This is called a double hysteresis curve.
2.3.2 PHENOMENOLOGY OF ANTIFERROELECTRICS

We will discuss here the introduction of electrostrictive coupling in Kittel’s free energy expres-
sion for antiferroelectrics.6,7 The simplest model for antiferroelectrics is the “one-dimensional two-
sublattice model.” It treats the coordinates as one-dimensional, and a superlattice (twice the unit
lattice) is formed from two neighboring sublattices each having a sublattice polarization Pa and Pb.
The state Pa = Pb represents the ferroelectric phase, while Pa = −Pb, the antiferroelectric phase. For
the electrostrictive effect, ignoring the coupling between the two sublattices, the strains from the
two sublattices are QPa2 and QPb2, respectively (assuming equal electrostrictive constants Q for both
sublattices). The total strain of the crystal becomes
(
x = Q Pa2 + Pb2 /2.) (2.62)
However, since antiferroelectricity originates from the coupling between the sublattices, it is appro-
priate to consider the sublattice coupling also for the electrostrictive effect. The coupling term for
the electrostriction Ω is introduced in the following form:
( ) ( )
G1 = (1/4)α Pa2 + Pb2 + (1/8)β Pa4 + Pb4 + (1/12) γ Pa6 + Pb6 ( )
+ (1/2)ηPa Pb − (1/2 )χ T p2 + (1/2)Qh ( Pa2 + Pb2 + 2Ω Pa Pb ) p , (2.63)

in which hydrostatic pressure p is employed, and χT is the isothermal compressibility, Qh (=Q11 + 2Q12)
and Ω are the electrostrictive constants. Introducing the transformations PF = (Pa + Pb)/2 and PA =
(Pa − Pb)/2 leads to the following expression:
(
G1 = (1/2)α PF2 + PA2 ) + (1/4)β (P F
4
+ PA4 + 6PF2 PA2 )
( )
+ (1/6 ) γ PF6 + PA6 + 15 PF4 PA2 + 15PF2 PA4 + (1/2)η PF2 − PA2 ( )
(
− (1/2 )χ T p2 + Qh⎡ PF2 + PA2 + Ω PF2 − PA2
⎣ )⎤⎦ p . (2.64)
The dielectric and elastic equations of state follow as
∂G1/∂PF = E = PF [α + η + 2Qh (1 + Ω) p + β PF2 + 3βPA2 + γPF4
+ 10 γPF2 PA2 + 5γPA4 ] (2.65)
∂G1/∂PA = 0 = PA [α − η + 2Qh (1 − Ω) p + β PA2 + 3bPF2 + γPA4
+ 10 γPF2 PA2 + 5γPF4 ] (2.66)
∂G1 /∂p = ΔV /V = −χ T p + Qh (1 + Ω)PF2 + Qh (1 − Ω)PA2 . (2.67)
Note that the macroscopic E for an antiferroelectric polarization configuration should be zero
(Equation 2.66). Hence, the induced volume change in the paraelectric phase can be related to the
induced ferroelectric polarization by the following formula:
( Δ V/V )ind = Qh (1 + Ω)PF,ind

2
. (2.68)
Below the phase transition temperature (this temperature for antiferroelectrics is called Neel tempera-
ture) the spontaneous volume strain and the spontaneous antiferroelectric polarization are related as
( ΔV /V )S = Qh (1 − Ω)PA,S
2
. (2.69)
Even if the perovskite cystal shows Qh > 0, the spontaneous volume strain can be positive or negative
depending on the value of Ω (Ω < 1 or Ω > 1), that is, if the intersublattice coupling is stronger than the
intrasublattice coupling (i.e., Ω > 1), a volume contraction is observed at the Neel point. This is quite
different from ferroelectrics, which always show a volume expansion at the Curie point. Figure 2.10
illustrates the spontaneous strains in a crystal schematically for Ω > 0. When Pa and Pb are in the par-
allel configuration (ferroelectric phase), the Ω term acts to increase the strain xS; when they are in the
antiparallel configuration (antiferroelctric phase), the Ω term acts to decrease the strain.
This phenomenological theory explains well the experimental results for the antiferroelectric
perovskite crystal PbZrO3 and others.8 Figure 2.11 shows the strain in the antiferroelectric ceramic
Pb0.99 Nb0.02[(Zr0.6Sn0.4)0.94 Ti0.06]0.98O3 as a function of an applied electric field.9 The large change
in the strain associated with the field-induced transition from the antiferroelectric to ferroelectric
phase can be estimated to be
( ΔV/V ) = Qh (1 + Ω)PF,S
2
+ Qh (1 − Ω)PA,S
2
= 2Qh ΩPF,S
2
. (2.70)
Here, we assume that the magnitudes of Pa and Pb do not change drastically through the phase
transition; that is, PF,S ≈ PA,S.

(a)
x = Q Pa2 x = Q Pb2
(b)
FIGURE 2.10 Intuitive explanation of the sublattice coupling with respect to electrostriction (for Ω > 0).
Ferroelectric arrangement: X = Q(1 + Ω) (Pa + Pb)2/4. Antiferroelectric arrangement: X = Q(1 – Ω) (Pa + Pb)2/4.
Ferroelectric phase
Strain (×10–4)
–4 –2 0 2 4
Electric field
Antiferroelectric phase (kV/mm)
FIGURE 2.11 Field-induced strain in a Pb(Zr,Sn)O3-based antiferroelectric.
2.4 FERROELECTRIC DOMAIN CONTRIBUTIONS

2.4.1 FERROELECTRIC DOMAIN REORIENTATION
In the phenomenological theory discussed above, it was assumed that the object materials are
monodomain single crystals and the application of an electric field does not change their state.
However, in ordinary piezoelectric ceramics, these assumptions do not strictly hold. An actual
device material sometimes shows a multiple domain structure even in a single crystal form, and
a much more complicated configuration in a polycrystalline ceramic. Figure 2.12 shows the elec-
tric field-induced strains, one parallel and the other perpendicular to the field, in a composition

–3
×10
x3
e
Δl/l
4
d
3 c
b
a
–2 –1 0 1 2
E (kV/mm)
x1 –1
FIGURE 2.12 Electric field-induced strain in a piezoelectric ceramic PLZT 7/62/38. From a to e, the maxi-
mum electric field increases.
7/62/38 of the well-known piezoelectric ceramics (Pb,La)(Zr,Ti)O3 (PLZT). In a cycle with a small
maximum electric field, the field-induced strain curve appears nearly linear, and can be called the
“inverse piezoelectric effect.” However, with the increase of the maximum electric field, the hys-
teresis becomes larger, and finally transforms to a symmetric butterfly shape when the electric field
exceeds a certain critical value (this is called coercive field). This is caused by the change of polar-
ity in ferroelectric domains under applied electric field, inducing a different state of polarization.
Strictly speaking, this PLZT includes two stages to be taken into account:
1. Domain reorientations in each grain

2. The polycrystalline state (a complex of randomly oriented tiny crystals)
Figure 2.13 shows the domain reorientation observed in BaTiO3 single crystals under an exter-
nal field. The 90° domain walls disappear independent of each other with increasing electric field
–700 V
–700 V
–700 V
700 V
700 V
700 V
0V
0V
0V
(a) (b) (c)
FIGURE 2.13 Domain reorientation process in a BaTiO3 single crystal. E direction: ® (leftward on the
figure). Around 300 V a most complicated domain pattern happens (a), which corresponds to the domain rever-
sal at the coercive electric field. Increasing the field (b) and (c), monodomain state is realized.

Strain x
4 3
1. E = 0 2. E = Ec 1
5
2
Field E
3. E = Emax
4. E = Ec
5. E = 0
E
FIGURE 2.14 Schematic depiction of the strain change in a ferroelectric ceramic associated with the
polarization reorientation.
finally to become a monodomain state throughout the entire crystal. However, in a polycrystalline
specimen, the domain wall motion is suppressed by the grain boundary, and a purely monodomain
state cannot be achieved.
Figure 2.14 shows schematically the domain reorientation in a polycrystalline body. Suppose a
sample is initially poled along the negative direction, and an external electric field is applied in the
positive direction, the crystal should shrink at first with an increase of the field because the field is
opposite in direction to the remnant polarization. Then, the deformation reaches a minimum at a
certain field EC (coercive field), where the polarization reversal begins in each grain. Above EC the
crystal starts to expand until E = Emax. Near Emax, all the reversible polarizations have been reversed,
and the crystal displays “piezoelectric” behavior once again with small hysteresis. In the process
of decreasing the electric field, there is no need for polarization reversal except in some unstable
domains with internal stresses. Until the field is reduced to zero, the strain decreases monotonically.
The final state at E = 0 is equivalent to the initial state but with the polarization direction reversed.
It can be said that the crystal is poled in the positive direction. It is interesting to point out that the
ratio σ of the transverse contraction strain to the longitudinal expansion strain (Poisson’s ratio) has
similar values for the perovskite-type piezoelectric ceramics, and is about 0.3.
EXAMPLE PROBLEM 2.9

BaTiO3 exhibits a rhombohedral crystal symmetry at liquid nitrogen temperature (i.e., −196°C)
and the angle distorted from the cubic structure is not very large (about 1°). Calculate all the
possible angles between the two non-180° domain walls. Neglect the angle difference less than
1° in the problem.
Solution
The polarization direction of BT at the low temperature is in the [111] or its equivalent axis of
a perovskite cell. Let us consider three typical directions [111], [111̄], and [11̄1̄] as illustrated in
Figure 2.15. The domain wall plane must be normal to one of the following and their equivalent
axes:

3 (111)
1
(111)
(111)
FIGURE 2.15 Polarization directions of the rhombohedral BT.
⎡⎣111⎤⎦ − [11 1] = [002]
⎣⎡111⎤⎦ − [111] = [022] (P2.9.1)
Thus, the angle between two of the non-180° domain walls is calculated as follows:
1. (002)/(200), (022)/(022), (002) /(220)

(002)⋅ (200) = 0
= 2 ⋅ 2 cos θ, θ = 90°.
2. (022)/(220), (022)/(220)
(022)⋅ (220) = 4 or − 4
= 2 2 ⋅ 2 2 cos θ, θ = 60° or 120°.
3. (002)/(022), (002) / (022)

(002)⋅ (022) = 4 or − 4
= 2 ⋅ 2 2 cos θ, θ = 45° or 135°.
In the following subsections, the Uchida–Ikeda theory which treats domain reorientation in
ceramics quantitatively, and the relationship between the crystal structure and the coercive field
are introduced.
2.4.2 UCHIDA–IKEDA MODEL

Let us take an example, BT, which has a tetragonal symmetry at room temperature. X-ray diffrac-
tion indicates a slight elongation along the [001] direction of a cubic perovskite cell with c/a = 1.01.
Therefore, when an electric field is applied on a monodomain single crystal along the a-axis, the 90°
domain reorientation from an a to c domain is induced, subsequently leading to a strain of 1% along
the field direction. However, the situation is much more complicated in the case of a polycrystalline
sample. Uchida and Ikeda treated this problem statistically, assuming grains (small crystallites)
being randomly oriented crystallographically.10,11
In an unpoled homogeneous polycrystalline ceramic, there must be no remanent polarization.
Let this state be the basis for zero strain. If an electric field E3 (in the 3-direction) is applied on

this sample, a polarization P3 will be induced, also the strains x1, x2, and x3, where x1 = x2 = −σx3
(σ: Poisson’s ratio). Let the spontaneous polarization and the principal strain of the individual crys-
tallites be PS and SS, respectively. For uniaxial crystals such as those with tetragonal and rhombohedral
symmetries, the principal strain SS is given in the direction of P3:
SS = (c/a ) − 1 (tetragonal crystal) (2.71)
SS = (3/2)(π /2 − α ) = (3/2) δ (rhombohedral crystal). (2.72)
A remark should be made here on the distinction between the principal strain and the spontaneous
strain (there was some confusion even in the original papers). Taking BaTiO3 as an example, the
principal strain is SS = 0.01, but the spontaneous strains should be defined as
x3,S = (c /a0 ) − 1,
x1,S = x2,S = (a /a0 )− 1, (2.73)
where a 0 is the lattice parameter of the paraelectric phase, and have the values 0.0075 and −0.0025,
respectively.
First, letting θ be the angle between the direction of the spontaneous polarization PS of a micro-
scopic volume Δv in a ceramic and the direction of the electric field E3, then the polarization P3 is
given by
∫
P3 = PS cos θ dv
∫ dv = P cos θ,
S (2.74)
where cos θ is the average value of cos θ in all the microscopic volumes of the ceramic. The
average strain is also calculated from the orientation of the strain ellipsoid:
∫ cos θ dv ∫ d v ⎠⎞ − 1/3⎥⎦ = S (cos θ − 1/3).

⎡ ⎤
x3 = SS ⎢⎛ 2 2
(2.75)
⎣⎝
S
This model, in which the microscopic regions with spontaneous strain change only their orienta-
tions, accompanied by no volume change, provides σ = 0.5 and
x1 = x2 = − x3 /2. (2.76)
However, there is a serious discrepancy with experimental data.

Next, in order to find the trend for the change in induced strain with an applied electric field, the
relationship between θ and E3 has to be known. Uchida et al. analyzed this problem by introducing
a characteristic angle θ90 for non-180° domain reorientations; that is, 90° reorientation in tetragonal
crystals, and 71° and 109° reorientations in rhombohedral crystals. But in order to simplify the
explanation, all reorientations are being represented by the former. Suppose a 90° domain rotation
occurs in a small region dv in a ceramic, and as a result, the polarization orientation of dv becomes
θ (angle from E3). These authors assumed that there exists a characteristic angle θ90, such that if
θ < θ90, a 90° rotation of the small region can occur, whereas if θ > θ90, no rotation will occur and
the region remains in its initial state. (This is a simple 1 or 0 model, which can be extended to a
probability model.) For a given θ90, which corresponds to a certain E3, by integrating Equation 2.75
over a volume dv for which θ < θ90 is satisfied, the induced strains x3, x1, and x2 can be obtained as
a function of θ90. Figure 2.16 shows the relationship between θ90 and (cos2 θ − 1/3), and Figure 2.17a
shows the measured values of induced strain in rhombohedral PZT ceramics. Comparing the two

0.003 0.424
Rhomb
0.4 0.002 0.368
0.001
Tet
0.3 0
cos2 θ – 1/3
0 1 2 3
0.2 Rhombohedral
Tetragonal
0.1
1 π 2
sin–1 sin–1 √
√3 4 3
0
0 10 20 30 40 50 60
θ90(°)
( )
FIGURE 2.16 Relation between θ90 and cos2 θ − 1/3 , where θ90 is a critical angle relating to the non-180°
( )
domain reorientation and cos2 θ − 1/3 is proportional to the field-induced strain.
×10–5
120
x1
30
Induced strain
80 20
θ90
40 10
0 0
–40 –20 0 20 40 –40 –20 0 20 40
(a) Electric field E3 (kV/cm) (b) Electric field E3 (kV/cm)
FIGURE 2.17 Transverse strain x1 vs. field in Pb(Zr0.57Ti0.43)O3, (a) and the calculated θ90 –E3 relation (b).
The measurement was done at 30°C.
θ90 and E3 figures reveals the relationship between θ90 and E3 (Figure 2.17b). It is apparent that
pronounced hysteresis also appears in the θ90 vs. E3 curve.
Furthermore, by finding the polarization P3 and the field-induced strain x3 (or x1) as a function
of the electric field E3, it is possible to estimate the volume in which a 180° reversal or a 90° rota-
tion occurred. This is because the 180° domain reversal does not contribute to the induced strain,
only the 90° rotation does, whereas the 180° domain reversal contributes mainly to the polarization.
It is shown schematically in Figure 2.18 that with the application of an electric field the 180° reversal
occurs rapidly whereas the 90° rotation occurs slowly.12 It is notable that at G in the figure, there
remains some polarization while the induced strain is zero, at H the polarizations from the 180° and
90° reorientations cancel each other and become zero, but the strain is not at its minimum. Generally
in such a case, a plot of the induced strain x3 vs. polarization P3 shows large hysteresis, exemplified
in a tetragonal PLZT (Figure 2.19).13 However, for materials whose polarization is dominated by
non-180° domain rotations, the hysteresis in the x vs. P plot should hardly be observed. Such is the
case for the low temperature phase of Pb(Mg1/3Nb2/3)O3, which is shown in Figure 2.20b.14

180°
Electric field
K J I H G (F, A)B C D E
90°
Reorientation
FIGURE 2.18 Electric field dependence of the domain volume fraction of 180° reversal (a) and of 90° reori-
entation (b). Notice the deviation of the zero fraction points, I and G, between 180° and 90°.
x3 × 103
2
1.5
0.5
Pn = P3/P3 max
–1 –0.5 0 0.5 1
FIGURE 2.19 P3–x3 relation in a tetragonal PLZT ceramic (6.25/50/50).

P3 (c/m2)
0.2
P3 (C/m2)
–0.2 –0.1 0 0.1 0.2
0.1 P3–E3
E3(V/m)
–1 1×106 –1
x1
–1
–2×10–4
x1
x1–E3
–2×10–4
(a) (b)
FIGURE 2.20 (a) P3–E3, x1–E3, (b) P3–x1 relations in a rhombohedral crystal of Pb(Mg1/3Nb2/3)O3
at −110°C.
Finally, the saturation values of polarization and strain of a ceramic under high electric field are
summarized:
Tetragonal: P3 → 0.831PS x3 → 0.368SS

Rhombohedral: P3 → 0.861PS x3 → 0.424SS

TABLE 2.3
Principal Strain, Spontaneous Polarization, Reoriented Volume Fraction, and Coercive
Field in Tetragonal and Rhombohedral PLZT Ceramics
Spontaneous Oriented Coercive
Crystal PLZT Principal Polarization Volume Field Calculated
Symmetry Specimen Strain Ss (%) (mC/cm2) Fraction (%) Ec (kV/cm) Ec (kV/cm)
Tetragonal 25/50/50 2.4 71 22 18 17.8
25/52/48 2.2 72 28 14.7 18.8
5/50/50 2.16 65 18 16.3 13
5/52/48 1.96 64.5 23 14.8 13.7
5/54/46 1.68 65 30 11.7 13
Rhombohedral 25/58/42 0.732 56.5 86.5 8.2 7
25/60/40 0.74 58.5 78.5 7.6 5.4
6/65/35 0.65 45 85 5.6 5.9
6.25/60/40 0.61 49 85 5.7 4.8
2.4.3 CRYSTAL STRUCTURE AND COERCIVE FIELD

In the previous subsection, a comparison was made between the tetragonal and rhombohedral sys-
tems from the viewpoint of saturation values for ceramics under a sufficiently large electric field. In
this subsection the difference between the two systems is illustrated with respect to the more useful
quantity, the coercive field.
Table 2.3 summarizes the principal strain SS, spontaneous polarization PS, volume percent of
reoriented domains γ90, and coercive field EC of PLZT ceramics which were reported by Schmidt.13
What can be understood from this table is that in comparison with the compositions with tetragonal
symmetry, those with rhombohedral symmetry have smaller principal strains, hence easier domain
rotation (larger γ90) and smaller coercive field EC. An expression for EC has been derived13,15
EC = αYSS2 γ 90 /PS , (2.77)
in which Y is the Young’s modulus, and α is a factor which takes into account the difference in
domain orientation between neighboring grains, and has values of 0.1 and 0.074 for the tetragonal
and rhombohedral systems, respectively.
2.4.4 HYSTERESIS ESTIMATION PROGRAM

On using a statistical treatment of the domain reversal mechanism, there are various computer
simulation trials in polarization (P–E) curves. We review here the calculation algorithm proposed
by R. C. Smith et al.16 Smith et al. adopted the so-called mesoscopic model to treat ferroelectric
materials. Let us consider primarily a ferroelectric such as PbTiO3, the crystallographic structure of
which is schematically illustrated in Figure 2.21a. There are basically two potential minima for the
Ti ion position, so that we can assume the double well model in Figure 2.21b.
Introducing the Helmholtz energy (temperature stability of polarization) for this double-well
polarization model at a finite temperature T
ψ ( P, T ) = U − ST
Φ0 N TkN ⎡ ⎛ P + PS ⎞ ⎤
= ⎡1 − ( P /PS )2 ⎤ + ⎢ P ln ⎜ ⎟ + PS ln(1 − ( P /PS )2 ⎥ (2.78)
4V ⎣ ⎦ 2VP
S ⎣ ⎝ PS − P ⎠ ⎦

x3
Pb
Internal energy
Ti P0
0 Ti position
(a) (b)
FIGURE 2.21 (a) PbTiO3 crystal lattice model. There are two potential minima for the Ti ion position.
(b) Double well model, showing the two potential minima. (From Smith, R.C. et al., J. Intell. Mater. Syst.
Struct., 14, 719, 2003. With permission. Smith, R.C., Smart Material Systems: Model Development, SIAM,
Philadelphia, PA, 2005.)
we obtain the following relationship
⎧ η
⎪ ( P ± PR )2 , P ≥ PI
⎪ 2
ψ(P) = ⎨ . (2.79)
⎡ 2 ⎤
⎪ η ( PI − PR ) ⎢ P − PR ⎥ , P < PI
⎪⎩ 2 ⎣ PI ⎦
Then, we introduce the Gibbs energy for a ferroelectric
G( E, P, T ) = ψ ( P, T ) − EP (Ferroelectric). (2.80)
The Gibbs energy curves for various electric fields are depicted in Figure 2.22.
In order to introduce the temperature fluctuation of the polarization more explicitly, Boltzmann
probability is introduced (refer to Figure 2.23):
kT e − G ( E ,P0 (T ),T )V /kT

μ(G ) = Ce − GV /kT ⇒ p+− = ⋅ ∞ . (2.81)
2πmV 2 /3
∫
e − G ( E ,P ,T )V /kT dP
P0
Ψ(P) = G(0, P)
G(E1, P) G(E2, P)
P P P
P0 PI PR
P P PR P
PI
E E E
Ec
FIGURE 2.22 Gibbs energy curves for various electric field E levels. (From Smith, R.C. et al., J. Intell.
Mater. Syst. Struct., 14, 719, 2003. With permission. Smith, R.C., Smart Material Systems: Model Development,
SIAM, Philadelphia, PA, 2005.)

G = Ψ – EP
PR P
–PI
PI
P
EC E
FIGURE 2.23 Introduction of Boltzmann probability to realize the rounded hysteresis curve in the polar-
ization and electric field relation. (From Smith, R.C. et al., J. Intell. Mater. Syst. Struct., 14, 719, 2003. With
permission. Smith, R.C., Smart Material Systems: Model Development, SIAM, Philadelphia, PA, 2005.)
By taking the evolution relations:
d x+ d x−
= − p+− x+ + p−+ x− , = − p−+ x− + p+− x+ (2.82)
dt dt
we finally obtain the polarization under a uniform lattice.
P = x+ P+ + x− P− , (2.83)
where
P+ =
∫ Pμ(G)dP.
P0
(2.84)
Figure 2.24 demonstrates the P–E hysteresis curves for PZT 5A piezoelectric. Notice that the hys-
teresis curves starting from any intermediate external parameter E exhibit a reasonable agreement
with the experimental curves.
0.4
Data
0.3 Model
0.2
Polarization (C/m2)
0.1
–0.1
–0.2
–0.3
–0.4
–3 –2 –1 0 1 2 3
Electric field (MV/m)
FIGURE 2.24 Hysteresis curves obtained from the mesoscopic approach in comparison with the experimen-
tal data: polarization vs. electric field in piezoelectric PZT 5A. (From Smith, R.C. et al., J. Intell. Mater. Syst.
Struct., 14, 719, 2003. With permission. Smith, R.C., Smart Material Systems: Model Development, SIAM,
Philadelphia, PA, 2005.)

2.4.5 DOMAIN ENGINEERING

In 1981, Kuwata et al. first reported an enormously large electromechanical coupling factor k33 =
92%–95% and piezoelectric constant d33 = 1500 pC/N in solid solution single crystals between
relaxor and normal ferroelectrics, Pb(Zn1/3Nb2/3)O3–PbTiO3.17,18 Figure 2.25a shows the piezoelec-
tric properties in the PZN–PT system, which has the morphotropic phase boundary (MPB) around
9 mol% of PT between the rhombohedral and tetragonal phases. The MPB composition exhibited
the giant piezoelectric d33 constant and electromechanical coupling factor k33eff only when the sam-
ple was prepared along [001] direction. This discovery has not been marked practically for more
than 10 years until high-k materials have been paid attention in medical acoustics.
These data have been reconfirmed by Yamashita et al. and Park et al., and improved data were
obtained recently, aiming at medical acoustic applications.19,20 The strains as large as 1.7% can be
induced practically for the PZN–PT solid solution single crystals, as shown in Figure 2.25b.20 It is
notable that the highest values are observed for a rhombohedral composition only when the single
crystal is poled along the perovskite [001] axis, not along the [111] spontaneous polarization axis. An
intuitive explanation for this enhancement of piezoelectricity is schematically shown in Figure 2.26.
Supposing that an electric field is applied on a single crystal with a cant angle from the spontaneous
dij (×10–10 C/N1) λ3 (×10–4 C/m2/k)
10
5 : Rhombo. :Tetr.
15 d*33
10
d*31
d33 d33 1.4
5
–d31 Single crystal
0 PZN–4.5% PT
1.0 1.2 (001)
k*33
k33
k33, –k31
–k31
0.5 k33 1.0 Single crystal
–k*31 PZN–8% PT Single crystal Single crystal
–k31 (001) PZN PMN–24% PT
0 (001) (001)
0.8
Strain (%)
sE33*
sEij (×10–10 m2/N)
1 0.6
sE33 sE33
sE11
0 0.4
4
εT3*
εT3 (×103)
E//(111) E//(001) 0.2 Ceramics, PZT-5H

2 Ceramics, PMN-PT
Ceramics, PZT-8
0
0 0.1 0.2 0.0
PZN x PT 0 30 60 90 120 150
(a) (b) Electric field (kV/cm)
FIGURE 2.25 (a) Piezoelectric properties in the Pb(Zn1/3Nb2/3)O3–PbTiO3 system (From Kuwata, J. et al.,
Ferroelectrics, 37, 579, 1981.) and (b) the electric field-induced strain curve. (From Park, S.E. and Shrout T.R.,
Mat. Res. Innovt., 1, 20, 1997. With permission.)

d33
PS
E1 PS eff
Strain d33
PS
E1
E
E
E2 d15 PS
E2
FIGURE 2.26 Basic concept of the piezoelectric enhancement in a perovskite ferroelectric. Note the fact
d15 >> d33 > |d31|.
polarization direction, the field component parallel to the PS generates the extension and contraction
strains via d33 and d31, and the component perpendicular to the PS generate the shear strain via d15.
Because d15 is much larger than d33 or d31 in perovskite ferroelectrics, this large shear strain exhibits
a drastic enhancement in the effective strain magnitude, as illustrated in Figure 2.26.
Another epoch-making paper was published in 1998; a series of theoretical calculations made on
perovskite-type ferroelectric crystals suggested that large d and k values in similar magnitudes to
PZN–PT can also be expected in PZT. Crystal orientation dependence of piezoelectric properties
was phenomenologically calculated for compositions around the MPB of PZT (see Figure 2.27).21
The maximum longitudinal piezoelectric constant d33eff (4–5 times enhancement) and electrome-
chanical coupling factor keff
3 (more than 90%) in the rhombohedral composition were found to be
at around 57° angle, canted from the spontaneous polarization direction [111], which corresponds
roughly to the perovskite [100] axis. Figure 2.28 illustrates the domain configuration difference
between the [111]-oriented film (monodomain-like with PS along the 1-axis) and the [100]-oriented
film (uniformly distributed domain pattern with PS along the 1-, 2-, 3-, and 4-axes). This is a trig-
ger report on “domain engineering.”22 This paper also predicted the superior design of the PZT
epitaxially grown thick/thin films (in addition to single crystal forms); that is, the rhombohedral
composition film of [100] orientation shows lower EC, sharper polarization reversal, and more linear
induction of strain, in comparison with the film of [111] orientation (Figure 2.29).
(001)
Z
(001)
(100) (010)
(010)
(100)
X
Y
FIGURE 2.27 Crystal orientation dependence of the effective piezoelectric constant d eff
3 in PZT single crystal-
like samples. (a) PZT 40/60 tetragonal composition and (b) PZT 60/40 rhombohedral composition. (From Du,
X.H. et al., J. Appl. Phys. Lett., 72, 2421, 1998.)

2 4 1 3
71o 2 4
3 7 E3
8 8 6
5 6 7 5
(a) (b)
FIGURE 2.28 Domain configuration models for a rhombohedral PZT films/single crystals with (a) [111]
plate and (b) [100] plate forms. (From Du, X.H. et al., Jpn. J. Appl. Phys., 36, 5580, 1997.)
P ε33 = 296
Pr = 50 μC/cm2 P ε33 = 530
Pr = 50 μC/cm2
√3
0 E E
0
S
d33 = 71 (pC/N) d33 = 189 (pC/N)
E E
0 0
(a) (b)
FIGURE 2.29 Polarization and strain curves predicted for a rhombohedral PZT films/single crystals with
(a) [111] plate and (b) [100] plate forms. (From Du, X.H. et al., Jpn. J. Appl. Phys., 36, 5580, 1997.)
Figure 2.30 shows the comparison between the theoretical and experimental results for the
piezoelectric d31 and e31 constants in epitaxially grown PZT thin films.23 Note that the maximum e31
constant can be obtained in the rhombohedral phase near the MPB, and in [100] specimen (rather
than in [111] specimen). Damjanovic et al. reported similar supporting results.24
EXAMPLE PROBLEM 2.10

The crystal orientation dependencies of the dielectric constant ε3 and piezoelectric constant d33 of
a tetragonal PZT are schematically illustrated in Figure 2.31a and b. Let us consider a uniformly
oriented polycrystalline sample. Discuss the change in the ε3 and d33 before and after poling.
Solution
Before poling, because of a uniform crystalline distribution, the dielectric constant should have
an intermediate value between εmin and εmax, and the piezoelectric constant should be zero.

600
6.4
Rhomb. Tetra.
500 6
PZT(100) 5.6
400 (001)//
d 33 (pC/N)
5.2
2
31 C/m
(001)//
300 4.8
eeff
PZT(111)
eff
4.4
200 (111)//
Ceramic
4
Random orientation
100 Ceramic 3.6
(111)//
3.2
0 80/20 70/30 60/40 50/50 40/60
0.4 0.45 0.5 0.55 0.6
(a) PZ % PT PT (b) Zr/Ti content
FIGURE 2.30 Comparison between the theoretical (a) and experimental results (b) for the piezoelectric d31
and e31 constants in epitaxially grown PZT thin films. (From Kalpat, S. and Uchino, K., J. Appl. Phys., 90,
2703, 2001.)
(001)
Piezoelectric
constant dmax = 200
Permittivity
pC/N
(001)
εmax = 500
(100) (100)
(a) (b)
FIGURE 2.31 Crystal orientation dependencies of the (a) dielectric and (b) piezoelectric constants of a
tetragonal PZT.
Electric poling orients the polarization along the z-axis, thus, the permittivity approaches
εz, leading to a decrease in permittivity after poling. To the contrary, the piezoelectric constant
should increase monotonically with increasing poling field, finally exhibiting a saturation of d33
above a certain poling field (close to the coercive field).
Lab Demonstration 2.1 Domain Observation
Using an optical microscope, learn how to identify the domain configuration and polarization
directions on a single crystal ferroelectric, such as BaTiO3. Figure 2.32a shows an optical micro-
scope setup, and (b) an example domain picture in a BT single crystal. This BT sample has a square
plate with the perovskite [100] axis parallel to the crystal edges. Taking into account the models
discussed in Example Problem 2.9, identify the domain configuration at room temperature.
Comments
BT has tetragonal symmetry at room temperature. The polarization directions should be along
the single crystal edge orientations (i.e., six equivalent polarization directions). Thus, 45° cant
stripe patterns correspond to 90° domain walls between up-down PS and left-rightward PS as

700 V
0V
–700 V
(b)
FIGURE 2.32 (a) Optical microscope setup and (b) example domain picture in a BT single crystal.
schematically illustrated on Figure 2.32b. The head-to-tail configuration is required to satisfy the
charge neutrality (i.e., Gauss law, div P = 0), and the upward PS is from the electric field direction
(upward in the picture).
Chapter Essentials
1. Tensor representation: when two physical properties are represented using tensors of
p-rank and q-rank, the quantity which combines the two properties in a linear relation is
also represented by a tensor of (p + q)-rank.
2. A physical property measured along two different directions must be equal if these two
directions are crystallographically equivalent. This consideration reduces the number of
the independent tensor components representing the above property.
' =
d ijk ∑a a
l ,m,n
il jm akn dlmn (a , a
il jm , akn : transformation matrices)
d 'ijk = dijk
3. Shear strain: x5 = 2x31 = 2φ, taken as positive for smaller angle.

4. Phenomenology:
(1/4)βP4 > 0 → second-order phase transition
(1/4)βP4 < 0 → first-order phase transition
5. Electrostriction equation:
x = QPS2 +2ε0εQPSE +Qε02 ε2 E 2

spontaneousstrain piezostriction electrostriction
The electrostrictive Q constant is insensitive to temperature changes.

6. Piezoelectric equations:
x = s E X + dE
P = dX +ε0ε X E

7. In antiferroelectrics, consideration of sublattice couplings is essential to understand the sta-

ble sublattice polarization configuration and the significant jump in strain associated with
an antiferroelectric–ferroelectric phase transition induced by an external electric field.
8. Polarization and electric field-induced strain in a polycrystalline body:
Tetragonal: P3 → 0.831PS, and x3 → 0.368SS
Rhombohedral: P3 → 0.861PS, and x3 → 0.424SS
Coercive field: Tetragonal > Rhombohedral
9. Rhombohedal PZT composition with the [100] perovskite electric field application induces
the maximum piezoelectric performance.
Check Point
1. Elastic Gibbs energy is given by

G1 = (1/2)αP 2 + (1/4)βP 4 + (1/6)γP 6 +
−(1/2)sX 2 − QP 2 X .
Why do not we include the “odd-number” power terms of P? Answer simply.
2. (T/F) The phenomenology suggests that the permittivity of a ferroelectric material exhibits
the maximum at its Curie temperature. True or False?
3. (T/F) The phenomenology suggests that the piezoelectric constant of a ferroelectric mate-
rial exhibits the maximum at its Curie temperature. True or False?
4. (T/F) The phenomenology suggests that the Curie–Weiss temperature of a ferroelectric
material is always higher than (or equal to) the Curie temperature. True or False?
5. How is the piezoelectric coefficient d related to the electrostrictive coefficient Q, spontane-
ous polarization PS, and relative permittivity ε in the phenomenology? Provide a simplest
formula.
6. In a polycrystalline PZT, we observed PS = 27 μC/cm2. Estimate the single crystal value
roughly with the Uchida–Ikeda model.
7. (T/F) The following two force configurations are equivalent mathematically. True or False?
(Figure 2.33)
8. The piezoelectric dijk tensor is a third-rank tensor. How many independent components are
there in general?
9. Provide three independent piezoelectric tensor components for a 4 mm crystal using a
reduced notation dij.
10. A typical PZT shows d15 = 600 × 10−12 m/V. When an electric field 1 kV/mm is applied to this
PZT perpendicular (1-axis) to the polarization direction (3-axis), calculate the angle change of
this ceramic sample between 1- and 3-axes in degree angle. 0.0003°, 0.003°, 0.03°, 0.3°, or 3°?
FIGURE 2.33 Shear-related forces.

Chapter Problems
2.1 The room temperature form of quartz belongs to class 32.

1. Find the piezoelectric matrix (dij) given by
⎛ d11 − d11 0 d14 0 0 ⎞

⎜ 0 0 0 0 − d14 −2d11 ⎟ .
⎜ ⎟
⎝ 0 0 0 0 0 0 ⎠
Notice that the piezoelectric tensor must be invariant for a 120° rotation around the
3-axis and for a 180°rotation around the 1-axis. The transformation matrices are
⎛ 1 3 ⎞
⎜ −2 2
0⎟
⎜ ⎟ ⎛1 0 0⎞
⎜ ⎟ ⎜ 0 −1 0 ⎟ , respectively.
3 1
−
⎜ 2 − 0 ⎟ and
2 ⎜ ⎟
⎜ ⎟ ⎝0 0 −1⎠
⎜ 0 0 1 ⎟
⎜ ⎟
⎝ ⎠
2. The measured values of the dij for right-handed quartz are
⎛ −2.3 2.3 0 −0.67 0 0⎞
⎜ 0 0 0 0 0.67 4.6⎟ × 10 −12 [C/N ] .
⎜ ⎟
⎝ 0 0 0 0 0 0⎠
a. When a compressive stress of 1 kgf/cm2 is applied along the 1-axis of a quartz crystal,
find the polarization generated. (kgf = kilogram force= 9.8 N)
b. When an electric field of 100 V/cm is applied along the 1-axis, find the strains
generated.
Hint
Let us initially use the following equation for 180° rotation (with respect to the a-axis):
d 'ijk = ∑a a
l ,m,n
il a d
jm kn lmn (a , a
il jm , akn : transformation matrices)
Considering that a11 = 1, a22 = −1, a33 = −1, we obtain the following relationship:
⎛ d ' = d111 ' = −d112

d112 ' = −d113 ⎞
d113
⎜ 111 ⎟
1st layer (i = 1) ⎜ d ' = −d121 ' = d122
d122 ' = d123 ⎟
d123
⎜ 121 ⎟
⎜ d131
' = −d131 ' = d132
d132 ' = d133 ⎟
d133
⎝ ⎠
⎛ d ' = − d211 d '212 = d212 d '213 = d213 ⎞
⎜ ⎟
211
2nd layer (i = 2 ) ⎜ d '221 = d221 d '222 = − d222 d '223 = − d223 ⎟

⎜ ⎟
⎜⎝ d '231 = d131 d '232 = − d232 d '233 = − d233 ⎟⎠
⎛ d ' = − d311 d '312 = d312 d '313 = d313 ⎞

⎜ ⎟
311
3rd layer (i = 3) ⎜ d321' = d321 d '322 = − d322 d '323 = − d323 ⎟

⎜ ⎟
⎜⎝ d '331 = d331 ' = − d332
d332 d '333 = − d333 ⎟⎠

or
⎛ d111 0 0⎞
⎜ ⎟
1st layer (i = 1) ⎜ 0 d122 d123 ⎟
⎜ 0 d133 ⎟⎠
⎝ d123
⎛ 0 d212 d231 ⎞
⎜ ⎟
2nd layer (i = 2 ) ⎜ d212 0 0⎟
⎜ d231 0 ⎟⎠
⎝ 0
⎛ 0 d312 d331 ⎞
3rd layer (i = 3) ⎜d 0 0⎟ .
⎜ 312 ⎟
⎝ d331 0 0⎠
Next, 120° rotation is considered such that
a11 = −1/2, a12 = 3 /2, a21 = − 3 /2, a22 = −1/2, a33 = 1:
' =
d111 ∑a a
l ,m,n
a d
1l 1m 1n lmn
= ∑a a
m,n
a d
11 1m 1n 1mn + ∑a
m,n
a a d
12 1m 1n 2mn
= (−1/2 ) d111 + (−1/2 )( 3 /2)2 d122 + 2 (−1/2 )( 3 /2)2 d212

3
= − (1/8)d111 − (3/8) d122 − (3/4 )d212
≡ d111
' =
d122 ∑a a
m,n
a d
11 2m 2n 1mn + ∑a
m,n
a a d
12 2 m 2 n 2 mn
( )
= (−1/2 ) − 3 /2 d111 + (−1/2 ) d122 + 2( 3 /2)( − 3 /2) (−1/2 )d212
2 3
= − (3/8)d111 − (1/8) d122 + (3/4 )d212
≡ d122 .
Continuing the calculations for d123, d133, d212, d231, d312, and d331 (total 8 components), we
can obtain all the necessary equations for deriving the final matrix form.
2.2 In the case of a first-order phase transition, the Landau free energy is expanded as in
Example Problem 2.6. Calculate the inverse permittivity in the vicinity of the Curie
temperature, and verify that the slope (∂(1/ε)/∂Τ) just below TC is 8 times larger than the
slope just above TC.
Hint
In the first-order phase transition, PS satisfies the following equation in the temperature
range of T < TC:

α + βPS2 + γPS4 = 0.
The permittivity is given by
1/ε0ε = α + 3βPS2 + 5γPS4 .
Thus
1/ε 0 ε = α + 3 β PS2 + 5( −α − βPS2 )
= −4α − 2βPS2.
Here α = (T − T0)/ε0 C, and PS2 = [ −β + β2 − 4 γ (T − T0 )/ε 0C ]/2γ .

It is important that the expansion series approximation should be obtained around the
temperature TC, not around the temperature T0; that is, (TC – T) is small, but (T0 − T) is
not a small number. Thus, from the following equation:
(T – T0 )/ε0C = (3/16)(β2 /γ ) − (TC − T )/ε0C,

we can obtain
1/ε 0 ε = −4α − 2βPS2
= −4[(3/16 )(β2 /γ ) − (TC − T )/ε 0C ]
⎡⎛ 3 ⎞ (β)2 ⎞ TC − T ⎤
+(β2 / γ ) − (β /γ ) β2 − 4 γ ⎢⎜ ⎟ ⎟− εC ⎥
⎣⎢⎝ 16 ⎠ γ ⎠ 0 ⎦⎥
Considering (TC − T) << 1, obtain the approximation of this equation. Two additional
hints: (1) 1 − x ≈ 1 − (x / 2 ) , and (2) β = −β when β < 0 .
2
2.3 In Figure 2.5, we introduced a critical temperature T1, above which the free energy will
diminish the potential minima related to the spontaneous polarization. Verify the dif-
ference between the temperature T1 and Curie–Weiss T0 temperatures expressed by
TC = T0 + (1/4 )(β2ε0C / γ )
for a first-order phase transition, where the Landau free energy is expanded as
F (P, T ) = (1/2 )αP 2 + (1/4 )βP 4 + (1/6 ) γP 6 ,
α = (T − T0 )/ε0C.
Hint
The potential minima are obtained from the first derivative of the free energy = 0:
(∂F /∂P ) = E = αP + βP 3 + γP 5 = 0.
This equation is valid for any temperature below and above Curie temperature. If we
take the second derivative of the free energy, its sign, positive or negative, corresponds

to the minimum or the maximum of the free energy. Since the free energy curve will
lose the maximum or minimum at T = T1, we obtain another condition:
(∂ 2 F /∂P 2 ) = α + 3βP 2 + 5γP 4 = 0.
This equation is only valid at the temperature T1. From these two equations
α + βP 2 + γP 4 = 0 → γ P 4 = −α − βP 2
α + 3βP 2 + 5γP 4 = 0 → α + 3βP 2 + 5( −α − βP 2 ) = 0.
Thus
α + β( −2α /β) + γ ( −2α /β)2 = 0 → α = β2 /4γ = (T1 − T0 )/ε 0C .
2.4 BT exhibits tetragonal crystal symmetry at room temperature and the distortion from
the cubic structure is not very large (c/a = 1.01). Calculate all the possible angles between
the two non-180° domain walls.
Hint
- -
Polarization vectors are of the following six types: [100], [100], [010], [010], [001],
-
and [001].
REFERENCES
1. J. F. Nye: Physical Properties of Crystals, p. 123, 140, Oxford University Press, London (1972).
2. A. F. Devonshire: Adv. Phys., 3, 85 (1954).
3. H. F. Kay: Rep. Prog. Phys., 43, 230 (1955).
4. K. Uchino, S. Nomura, L. E. Cross, S. J. Jang, and R. E. Newnham: Jpn. J. Appl. Phys., 20, L367 (1981);
K. Uchino: Proc. Study Committee on Barium Titanate, XXXI-171-1067 (1983).
5. J. Kuwata, K. Uchino, and S. Nomura: Jpn. J. Appl. Phys., 19, 2099 (1980).
6. C. Kittel: Phys. Rev., 82, 729 (1951).
7. K. Uchino: Solid State Phys., 17, 371 (1982).
8. K. Uchino, L. E. Cross, R. E. Newnham, and S. Nomura: J. Appl. Phys., 52, 1455 (1981).
9. K. Uchino: Jpn. J. Appl. Phys., 24(Suppl. 24-2), 460 (1985).
10. N. Uchida and T. Ikeda: Jpn. J. Appl. Phys., 6, 1079 (1967).
11. N. Uchida: Rev. Electr. Commun. Lab., 16, 403 (1968).
12. N. Uchida and T. Ikeda: Jpn. J. Appl. Phys., 4, 867 (1965).
13. N. A. Schmidt: Ferroelectrics, 31, 105 (1981).
15. P. Gerthsen and G. Kruger: Ferroelectrics, 11, 489 (1976).
16. R.C. Smith, S. Seelecke, Z. Ounaies, and J. Smith: J. Intell. Mater. Syst. Struct., 14(11), 719 (2003); R.C.
Smith: Smart Material Systems: Model Development, SIAM, Philadelphia, PA (2005).
17. J. Kuwata, K. Uchino, and S. Nomura: Ferroelectrics, 37, 579 (1981).
19. K. Yanagiwawa, H. Kanai, and Y. Yamashita: Jpn. J. Appl. Phys., 34, 536 (1995).
20. S. E. Park and T. R. Shrout: Mat. Res. Innovt., 1, 20 (1997).
21. X. H. Du, J. Zheng, U. Belegundu, and K. Uchino: J. Appl. Phys. Lett., 72, 2421 (1998).
22. X. H. Du, U. Belegundu, and K. Uchino: Jpn. J. Appl. Phys., 36(9A), 5580–5587 (1997).
23. S. Kalpat and K. Uchino: J. Appl. Phys., 90(6), 2703–2710 (2001).
24. D. Damjanovic, D. V. Taylor, and N. Setter: Proc. Mater. Res. Soc., Symp. VIII (1999).

3 Designing with Materials
and Devices and
Fabrication Processes
In this chapter, the material designing, such as solid solution composition and doping, is introduced
first, followed by the fabrication processes. The fabrication of ferroelectric ceramic devices gener-
ally involves two steps: preparation of the ceramic powders and sintering of the shaped structures.
Wet chemical preparation methods are utilized for producing ceramic powders in order to ensure
reproducibility of the advanced characteristics of the devices. Alternative fabrication processes
include single crystal preparations and thin-/thick-film techniques. Popular device designs include
multilayers, bimorphs, and other composite types. Necessary basic knowledge of particle size and
film thickness effect on ferroelectricity will also be discussed in conjunction with micro/nanotech-
nologies in this chapter.
3.1 MATERIAL DESIGNING

3.1.1 COMPOSITION SELECTION
As discussed in Chapter 1, one strategy for designing ferroelectric materials is to select a material
with a suitable Curie temperature. For capacitor dielectrics, the peak dielectric constant around the
transition (Curie) temperature is utilized, while for memory applications, the material must be ferro-
electric at room temperature. A large temperature dependence of the spontaneous polarization below
TC is sought for pyroelectric sensors. On the other hand, piezoelectric materials are used as sen-
sors and actuators, where the TC should be much above room temperature. Electro-optic materials,
which are key components in displays and optical communication systems, use mostly the Kerr
effect, which requires TC lower than room temperature. Finally, for thermistor applications, barium
titanate (BT)–based semiconductive ferroelectric ceramics are used by setting their TC around the
temperature at which a positive temperature coefficient (PTC) is desired.
Let us consider an example for piezoelectric application. In general, Pb(Zr,Ti)O3 (PZT),
PbTiO3 (PT), (Pb,La)(Zr,Ti)O3 (PLZT), and PZT-based ternary ceramics are utilized for piezo-
electric applications. Their piezoelectric coefficients are summarized in the international data
book:
K. H. Hellwege et al.: Landolt–Bornstein, Group III, Vol. 11, Springer-Verlag, New York (1979).
Figure 3.1 shows the composition dependence of the permittivity and the electromechanical
coupling factor kp for the PZT system.1
If we do not have this sort of comprehensive experimental data, how can we estimate the val-
ues for the solid solutions? In general, physical properties of a solid solution between A and B,
(1 − x)A − xB, can be estimated by a phenomenological theory.2,3 Expanding the theories in
Chapter 2, the Gibbs elastic energy of a solid solution is assumed to be a linear combination of the
Gibbs elastic energy of each component:
75

1600 0.40
Permittivity εT33 /ε0(25°C, 1 MHz)
EM coupling factor kp (25°C)

1200 0.30
800 0.20
400 0.10
0 0
0 20 40 60 80 100 PbZrO2
PbTiO2 100 80 60 40 20 0
Composition (mol%)
FIGURE 3.1 Composition dependence of the permittivity and the electromechanical coupling factor kp in
the PZT system. (From Jaffe, B. et al., Piezoelectric Ceramics, Academic Press, New York, 1971.)
G1 ( P, X , T ) = (1/2)[(1 − x )α A + xα B ]P 2 + (1/4)[(1 − x )βA + xβB ]P 4

+ (1/6)[(1 − x )γ A + x γ B ]P 6
−(1/ 2)[(1 − x )sA + xsB ]X 2 − [(1 − x )QA + xQB ]P 2 X .
(α A = (T − T0,A ) /ε 0CA , α B = (T − T0,B )/ε 0CB ). (3.1)
The solution provides reasonable first-order estimates of the Curie temperature, spontaneous
polarization and strain, permittivity, piezoelectric constant, and electromechanical coupling. Abe
et al. reported a good example of theoretical fitting to experimental results for the solid solution
Pb(Zn1/3Nb2/3)O3–PbTiO3.3 By expanding Equation 3.1 into the three-dimensional (3-D) expression
with P1, P2, P3, X11, X33, X44, etc. for realizing cubic, tetragonal, and rhombohedral symmetry phases,
the Gibbs energy is calculated for a certain composition x and temperature T. Then, the lowest
energy phase is taken as the stable phase at that particular point (x, T). Figures 3.2 through 3.5 show
these fittings calculated on the basis of the data presented in Table 3.1. The transition temperature
from the tetragonal to cubic phase (Curie temperature) changes linearly with the composition x, as
the reader can imagine, while the transition temperature from the rhombohedral to tetragonal phase
is calculated by comparing G1 for both phases.
Frequently Asked Question 3.1
Why does the morphotrophic phase boundary (MPB) composition exhibit high ε and high d coef-
ficients, in comparison with the compositions inside the phase, as shown in Figure 3.1?
Answer
Figure 3.6 illustrates the potential minimum wells around the MPB composition (Top-right), in com-
parison with a tetragonal (Bottom-left) and a rhombohedral composition (Bottom-right) of perovskite
PZTs. In the tetragonal phase, six equivalent energy minimum wells are expected at discrete positions
along [100], [1̄00], [010], [01̄0], [001], and [001̄] directions; while in the rhombohedral phase, eight
equivalent energy minimum wells are expected along [111], [111̄], [11̄1], [1̄11], [11̄1̄], [1̄11̄], [1̄1̄1], and
[1̄1̄1̄] directions. Since we expect 14 (=6 + 8) equivalent energy minimum wells rather closely located
for the MPB composition, the energy gap between each potential well is rather shallow, leading to

Designing with Materials and Devices and Fabrication Processes 77
TABLE 3.1
Coefficients for PZN (Pb(Zn1/3Nb2/3)O3) and
PT (PbTiO3) Used in Calculating the Transition
Temperature and Lattice Parameters in (1 − x)
PZN–xPT
Constants PZN PT
T0 (°C) 130 478.8
C (105 °C) 4.7 1.5
ξ11 (107 m5/FC2) −13.7 −29.0
ξ12 (108 m5/FC2) −1.0 15.0
ζ111 (108 m9/FC4) 10.3 15.6
ζ112 (108 m9/FC4) 6.8 12.2
Q11 (10−2 m4/C2) 2.4 8.9
Q12 (10−2 m4/C2) −0.86 −2.6
Q44 (10−2 m4/C2) 1.6 6.75
ac (Å) 4.058 3.957
Cubic
200
200 Cubic
Transition temperature (ºC)
Transition temperature (ºC)
Tetragonal Tetragonal
100
100
0
Rhombohedral 0
Rhombohedral
–100 –100
0 0.1 0.2 0 0.1 0.2
(a) PZN x PT (b) PZN x PT
FIGURE 3.2 Phase diagram for (1 − x)Pb(Zn1/3Nb2/3)O3 – xPbTiO3: (a) calculated and (b) experimental.
easy ionic shift/movement under an electric field. In other words, high permittivity (ionic polarization
induced by an electric field) and high piezoelectric d constant (ionic displacement and bulk strain
induced by an electric field) are expected. Further, we can say that the phase transition between the
tetragonal and rhomboderal phases will be easy under an electric field application. We can simply
say that the MPB composition is dielectrically and elastically “compliant” under an electric field.
3.1.2 DOPANT EFFECTS ON FERROELECTRICITY

Small amounts of dopants sometimes drastically change the dielectric, electromechanical, and
electro-optic properties of ferroelectric ceramics. The PTC effect in semiconductor BT in Chapter 9
is one of the most notable examples of the effect of dopants on electrical properties, similar to
p-type or n-type semiconductor properties.

4.3 4.3
rh. rh.
a= b =c a= b =c
Lattice parameters (å) 4.2 4.2
Lattice parameters (å)

c
c
4.1 4.1
4.0 4.0
a=b a=b
3.9 3.9
Tetr. Tetr.
3.8 3.8
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
(a) PT x PZN (b) PT x PZN
FIGURE 3.3 Lattice parameters for (1 − x)Pb(Zn1/3Nb2/3)O3 – xPbTiO3: (a) calculated and (b) experimental.
90°
89° a
a
88°
c 90°
a
89.9° a
89.8° c
a
a
4.05
a
4.05
a
a
Rhombo. Tetr. Cub.
Rhombo. Tetr. Cub.
4.00 4.00
–100 0 100 200 –100 0 100 200
(a) Temperature (°C) (b) Temperature (°C)
FIGURE 3.4 Temperature dependence of lattice parameters for 0.91 PZN – 0.09 PT: (a) calculated and
(b) experimental.
In this section we consider the “hard” and “soft” piezoelectric PZT properties. Hard or soft
property is defined by the coercive electric field of the PZT sample. Figure 3.7 depicts soft and hard
properties in the strain curve. When the coercive field is higher than 1 kV/mm, the PZT is called
“hard”; when the coercive field is less than 100 V/mm, it is called “soft.” When the coercive field is
in between, it is called “semihard.” Note that both samples have the same base composition PZT
52/48 with only 1 at.% (atomic percent) different dopants.
The difference of piezoelectric properties is compared in Table 3.2 among PMN-based elec-
trostrictor, PZT-based soft and hard piezoelectrics.
3.1.2.1 Domain Wall Stability

First, the “soft” and “hard” characteristics are a reflection of the coercive field EC, in other words, the
stability of the domain walls. Consider the transient state of a 180° domain reversal, which reveals a
domain wall front with a head-to-head polarization configuration (see Figure 3.8). From Gauss’s law
div P = ρ, (ρ :charge density) (3.2)

30,000 30,000
0.9
x = 1.0 0.8
0.95
x = 1.0
Dielectric constant 0.7
Dielectric constant
20,000 0.6 20,000
0.9
0.8
0.7
0.6
10,000 10,000
0 0
0 100 200 300 400 0 100 200 300 400
(a) Temperature (°C) (b) Temperature (°C)
FIGURE 3.5 Temperature dependence of dielectric constant for xPZN – (1 − x)PT: (a) calculated and
(b) experimental.
Free energy (001)

T > TC
T = TC
T < TC
(110)
P MPB
(a)
Tetragonal Rhombohedral
(001) (001)
a
c
Ps Ps
a a
a
(110) (110)
Tetragonal Rhombohedral
FIGURE 3.6 Illustration of potential minimum wells around the MPB composition.
Strain Strain
Electric Electric
field field
(a) (b)
FIGURE 3.7 Soft (a) and hard (b) properties of PZT samples in the strain curve.

TABLE 3.2
Comparison among Electrostrictor, Soft, and Hard Piezoelectrics
Off-Resonance Resonance
d k Qm Application Application
Electrostrictor (PMN) High High Low High displacement Wideband
(Under DC bias)a No hysteresis
Soft piezo (PZT-5H) High High Low High displacement Heat generationa
(ΔL = dEL)
Hard piezo (PZT-8) Low Low High Low straina High AC displacement
(ΔL μ QmdEL)
a Demerit.
Ps –Ps Ps –Ps
–
Easy to move Stable
(a) E (b) E
FIGURE 3.8 Domain wall stability in consideration of Gauss’s law. (a) Without space charge, and (b) with
movable charge in the crystal.
the domain wall front is very unstable in a highly insulating material (Figure 3.8a with no space
charge), leading to quick disappearance of this domain wall, i.e., a low coercive field. However,
if the material has movable space charges (Figure 3.8b) in the crystal lattice, the head-to-head
polarization configuration is stabilized, leading to a high coercive field. This is called domain wall
pinning model.
3.1.2.2 Crystallographic Deficiencies

We consider possible crystallographic deficiencies due to impurity doping in a perovskite crystal.
Acceptor ions, such as Fe3+, introduce oxygen deficiencies (❑) in the PZT lattice as follows:
Pb (Zry Ti1− y − x Fe x )(O3 − x/ 2 ❑ x /2).
This acceptor doping causes the easy reorientation of deficiency-related dipoles. These dipoles
are generated by an Fe3+ ion (effectively negative charge) and an oxygen vacancy site (effectively
positive charge). The deficiencies are produced at a high temperature (>1000°C) during sintering;
however, the oxygen ions are still movable even below the Curie temperature (e.g., at room tempera-
ture), because the oxygen and the vacancy are adjacent (only 2.8 Å!) and hopping occurs easily (see
Figure 3.9a). The movable oxygen ion is considered to stabilize the domain motion, i.e., “hard” PZT.
In the case of donor ions, e.g., Nb5+, Pb deficiency is introduced because of the charge neutrality
requirement in the crystal:
(Pb1− x /2 ❑ x /2 ) (Zry Ti1− y − x Nb x )O3.

Donor doping is not very effective for generating movable dipoles, because the Pb ion cannot easily hop
to an adjacent A-site vacancy due to the close oxygen (large ionic size) surroundings (Figure 3.9b).

O vacancy A vacancy
Easy to move Prohibited

3+
5+
(a) Acceptor doping (b) Donor doping
FIGURE 3.9 Crystal deficiencies in PZT for acceptor (a) and donor (b) dopants.
Moreover, it is noteworthy that lead-containing ceramics such as PZT tend to be p-type semi-
conductors due to Pb evaporation during sintering. Hence, donor-type ion doping is considered to
compensate the original acceptor-type or p-type deficiencies. Thus, donor-doped piezoceramics
facilitate the domain wall motion, i.e., “soft” PZT, exhibiting large piezoelectric d constants, and
also easy depoling under the usage.
The above “oxygen deficiency” and “domain-wall pinning” model is popularly used to explain
the origin of “hard” PZT. However, Uchino et al. proposed a different model to explain the “harden-
ing” mechanism of PZT, i.e., “internal bias” model originated from “oxygen diffusion,” as discussed
in the next section.
The previously mentioned crystal deficiency model helps us understand various ferroelectric
property changes with doping. Hagimura and Uchino 4 studied the effect of impurity doping on
field-induced strains using a base composition of (Pb0.73Ba0.27)(Zr 0.75Ti0.25)O3 (soft piezoelectric).
Figure 3.10a shows the definitions of the maximum strain xmax, and the degree of hysteresis
Δx/xmax (%). The degree of hysteresis is calculated from the strain deviation during the rise and
10
+4~+6
Group A
9 Rare earth
+1~+3
8
Maximum strain (1kV/mm)xmax (×10–4)
Th
7
Y
6
5 W Ta Undoped Group B
5 Emax 4 Nb
4 Emax/2 Mg Group C
Strain (×10–4)
3
3
Er
Δx
2 Pr Yb Cr
xmax
1 1 La Na Zn B
0 Fe
0 0.5 1.0
0 25 50
(a) Electric field (b) Hysteresis Δx/xmax (%)
FIGURE 3.10 Dopant effect on the field-induced strain. Maximum strain and hysteresis in (Pb0.73Ba0.27)
(Zr0.75Ti0.25)O3-based ceramics. (a) Definition of the maximum strain xmax and the degree of hysteresis Δx/xmax
(%), and (b) dopant effect on the actuator parameters.

fall of the field Δx at half the maximum electric field (1 kV/mm) divided by the maximum strain
xmax at 1 kV/mm. Figure 3.10b shows the dopant effect on the strain characteristics, in which the
doping concentration is fixed at 2 at.%. It is seen that materials incorporating donor-type ions
on the B site with a large valence +4 ~ +6 ion (Ta 5+, Nb5+, W6+) exhibit excellent characteris-
tics as positioning actuators; namely, enhancement of the strain magnitude and reduction of the
hysteresis.
To the contrary, the acceptor-type ions with a small valence +1 ~ +3 suppress the strain and
increase the hysteresis and the coercive field. Although acceptor-type ions must be avoided in design-
ing actuator ceramics used for positioner applications (off-resonance applications under high electric
field), acceptor doping is very necessary for preparing “hard” piezoceramics which are particularly
suitable for ultrasonic motor applications (resonance applications under lower electric field).
3.1.3 HIGH-POWER CHARACTERISTICS

The power density of a piezoelectric is measured by different figures of merit (FOMs) for different
applications:
1. Off-resonance actuator applications—positioners

FOM = d (piezoelectric constant)
2. Resonance actuator applications—ultrasonic motors
FOM = v (vibration velocity) ≈ Qm d or Qm k (for a low-level excitation)
3. Resonance transducer applications—piezoelectric transformers, sonars (transmitters and
receivers)
FOM = kv or Qm k 2 (k :electromechanical coupling factor)
In this section, we will focus on high-power piezoelectric ceramics for ultrasonic (AC drive) appli-
cations. In order to obtain a large output power, the ceramics are driven under a high vibration
level, namely under a relatively large AC electric field. This causes heat generation as well as a
drastic degradation in piezoelectric properties. Therefore, the high-power device such as an ultra-
sonic motor requires a very “hard” piezoelectric with a high mechanical quality factor Qm in order
to suppress heat generation. The Qm is defined as an inverse value of the elastic loss factor, tan φ′.
It is also notable that the actual mechanical vibration amplitude at the resonance frequency is
directly proportional to this Qm value (i.e., displacement amplification factor).
Since the measurement technique to determine the piezoelectric characteristics at high vibration
levels was not established previously, quite a few efforts were made recently initiated by Uchino
et al.5 and Hirose et al.,6 who have succeeded in developing a method for measuring the electro-
mechanical properties with higher precision and stability at high vibration levels with a constant
current circuit.
The vibration velocity dependence of the piezoelectric constant, permittivity, elastic compliance,
and electromechanical coupling factor was obtained for PZT-based samples.7 Figure 3.11 shows the
mechanical quality factor Qm and the actual temperature rise plotted as a function of the average
vibration velocity. Notice a significant decrease in Qm above a certain critical vibration level, which
is the limit of the output vibration energy. Even if the input electrical energy is increased further, the
additional energy will be converted only into heat. We define the maximum vibration velocity when
the temperature rise at the nodal point (maximum temperature point) reaches 20°C higher than
the room temperature. The temperature distribution in a rectangular plate (shown in Figure 3.11),
when the plate is driven at its first resonance frequency, is pictured in Figure 3.12. Note that the
highest temperature point corresponds to the nodal point where the maximum stress and strain
are realized. The difference between the A-type (resonance) and B-type (antiresonance) modes is
discussed in Chapter 7.

2000 QB 40
Temperature rise (ºC)

Quality factor Q
1500 QA 30
7
1000 2 20
43
Test sample
500 A-type 10
Temperature rise B-type
0 0
0.01 0.02 0.05 0.1 0.2 0.5 1.0
Vibration velocity v0 (m/s)
FIGURE 3.11 Vibration velocity (at the sample edge) dependence of the quality factor Qm and temperature
rise for both A (resonance) and B (antiresonance) type resonances of a longitudinally vibrating PZT ceramic
transducer through d31. The test sample size is shown in the figure.
23.00 62.25
FIGURE 3.12 Temperature distribution in a rectangular plate (shown in Figure 3.11), when the plate is driven
at its first resonance frequency.
60
Undoped Pb (Zr0.5 Ti0.4)O3
Temperature rise ΔT (°C)
0.5 wt% Nb-doped

40
0.5 wt% Fe-doped
20
0
0.01 0.05 0.1 0.5 1.0
Vibration velocity v0 (m/s)
FIGURE 3.13 Temperature rise ΔT vs. effective vibration velocity v for PZT samples.
Figure 3.13 shows the temperature rise vs. vibration velocity for undoped, Nb-doped, and
Fe-doped PZT samples. In comparison with the sample doped with Nb (donor) ion, which gener-
ates heat just around 0.05 m/s (rms value), the suppression of heat generation is significant in the
Fe(acceptor)-doped ceramic. Note that commercially available hard PZTs exhibit 0.3 m/s as the
maximum vibration velocity, vmax.
Figure 3.14 shows mechanical Qm vs. mole fraction of Zr (x) at effective vibration velocities
v0 = 0.05 m/s and 0.5 m/s for Pb(ZrxTi1 − x )O3 doped with 2.1 at.% of Fe3+.8 The decrease in mechanical
Qm with increase in vibration level is minimum around the rhombohedral–tetragonal MPB(52/48).
In other words, the worst material at a small vibration level becomes the best at a large vibration
level, and data obtained from a conventional impedance analyzer are not relevant to search high-
power materials.

Pb(ZrxTi1–x)O3 + 2.1 at.% Fe3+
200
Mechanical quality factor Q
Mechanical quality factor Q

Vibration velocity 60
v0 = 0.05 m/s
100 v0 = 0.5 m/s 40
20
0.48 0.50 0.52 0.54 0.56 0.58

Mole fraction of Zr
FIGURE 3.14 Mechanical quality factor Qm vs. mole fraction of Zr(x) at effective vibration velocities v0 =
0.05 and 0.5 m/s for Pb(ZrxTi1 − x)O3 + 2.1 at.% Fe3+ ceramics.
RA (directly measured)
= R d + Rm
100
Rd, Rm, and RA (Ω)
30
Rm = ωALA/QB
10
Rd Rd = tan δ/ωACd
3.0
LA
Cd CA
Rm
0.03 0.1 0.3 1.0
(a) (b) Vibration velocity v0 (m/s)
FIGURE 3.15 (a) Equivalent electrical circuit for the piezoelectric resonance mode, and (b) vibration veloc-
ity dependence of the resistances (damped Rd and motional Rm) in the equivalent electric circuit for a piezo-
electric component. Heat generation originates from a drastic increase of Rd with vibration velocity.
Figure 3.15 highlights the key material-related factors affecting heat generation in the piezoelec-
tric material. The resistances Rd (damped) and Rm (motional) in the equivalent electrical circuit
(Figure 3.15a) are separately plotted as a function of vibration velocity (Figure 3.15b).9 Note that
motional Rm, mainly related to the “mechanical” loss, is insensitive to the average vibration veloc-
ity, while damped Rd, related to the “dielectric” loss, changes significantly around a certain critical
vibration velocity. Thus, the resonance loss Qm−1 at a small vibration velocity is mainly determined
by the intrinsic mechanical loss, and with increasing vibration velocity, the intrinsic dielectric loss
contribution significantly increases. We can conclude that heat generation is caused primarily by
dielectric loss (i.e., P–E hysteresis loss). Refer to Section 7.3.2 for the equivalent circuit.

In order to provide a measure of vibration level, the vibration amplitude may be used. Why the
vibration velocity is used in the above discussion?
Answer
When the vibration amplitude is small and proportional to the applied electric field, it should be
expressed by
ΔL = (8/π 2 )Qmd31LE3 sin(ωRt ),
for a d31-type rectangular piezoelectric plate.

Since the vibration velocity is the first derivative of amplitude in terms of t, we obtain
v = (8/π 2 ) Qmd31LE3 ω R cos(ω Rt ).
Taking into account fR = 1 ( )

ρs11 2L , the vibration velocity can be transformed as
( ) (
v = 8 π 2 Qmd31L E3 2π 2L ρs11 cos (ωRt ))
( )
= (8 π ) Qmd31E3 1 ρs11 cos (ωRt ).
Note that the vibration amplitude is sample-size-dependent, but the vibration velocity is not.
Because the vibration velocity is provided primarily by Qmd31 ρs11 , which is a material’s con-
stant, we use it as a measure of the vibration level.
3.2 FABRICATION PROCESSES OF CERAMICS

Ferroelectric devices are typically fabricated from polycrystalline ceramics. This involves two
steps: preparation of the ceramic powders and sintering of the shaped structures. Single crystals are
expanding their applications because of superior performances in comparison with polycrystalline
forms. Template grain growth techniques are also focused on obtaining inexpensive single crystal-
like performance.
3.2.1 PREPARATION OF CERAMIC POWDERS

Particle shape, particle size distribution, and compositional uniformity are the key factors to be
controlled in the raw powder in order to realize reproducibility of the piezoelectric characteris-
tics. The usual method is the oxide-mixing technique, in which a desired chemical composition is
made by firing raw oxide powders (calcination), then crushing them into fine powders. Since the
oxide-mixing method results in difficulties in achieving microscopic compositional uniformity, wet
chemical methods (coprecipitation, alkoxide) have been employed more recently in manufacturing
ceramic devices. In this section, processes for fabricating BT, lead zirconate titanate (PZT), and
lead magnesium niobate (PMN) ceramics are reviewed.10
3.2.1.1 Solid-State Reaction

Let us consider the preparation of Pb(ZrxTi1 − x)O3 powders. The raw powders PbO, ZrO2, and TiO2
are weighed in an appropriate proportion, mixed, and calcined at around 800°C–900°C for 1–2 h.
Then the sample is crushed and milled into fine powders. The drawbacks here are that the milling
process does not efficiently give particles of size less than 1 μm, and that the contamination of the
sample by milling media is unavoidable. On the other hand, the accidental “Fe” contamination was

a trigger for discovering “hard” PZT; conventionally steel (Fe) balls were used for ball milling of
PZT powders.
BaTiO3 is prepared in principle from equimolar quantities of raw powders of BaO and TiO2. In
general, BaCO3 powder is recommended instead of BaO, because high-purity BaO is expensive and
chemically less reactive.
A similar calcination process starting from PbO, MgO, Nb2O5, and TiO2 can be used for
Pb[(Mg1/3Nb2/3)1−xTi x]O3. However, this simple process generates a second phase (pyrochlore) in
addition to the perovskite phase. To suppress this second phase several mol% excess PbO doped in
the final sintering stage is effective.11 Swartz et al. reported a unique method taking account of the
chemical reaction process.12 They demonstrated that the perfect perovskite phase can be obtained
by the reaction starting from columbite MgNb2O6 and PbO:
3PbO + MgNb 2O6 → 3Pb(Mg1/3Nb 2/3 )O3
For PMN–PT, MgO, Nb2O5, and TiO2 are mixed and fired at 1000°C initially. Then PbO is added
to the columbite, and the sample is calcined at 800°C–900°C. Several mol% excess MgO is particu-
larly effective in obtaining the perfect perovskite phase.
3.2.1.2 Coprecipitation
Since practical piezoelectric/electrostrictive ceramics are from complex solid solution systems, the
compositional fluctuation among ceramic particles as well as the purity is very important. The
above method involving a combination of solid-state reaction and mechanical mixing is inevitably
accompanied by this problem. The coprecipitation method is employed to realize enhanced homo-
geneity. Adding a precipitant into the liquid solution of mixed metal salts produces a homogeneous
precipitate. Thermal dissolution makes homogeneous powders from the precipitate.
For example, BaTiO(C2H4)2 . 4H2O with the Ba/Ti ratio exactly equal to 1:1, on an atomic scale,
can be precipitated from an aqueous solution of BaCl2 and TiCl4 by dripping in oxalic acid. Thermal
dissolution of this precipitate provides fine stoichiometric BaTiO3 powders with good sintering
characteristics.
For PLZT, Pb(NO3)2, La(NO3)3 . 6H2O, ZrO(NO3)2 . 2H2O, and TiO(NO3)2 are used as the starting
materials.13 First, the water solutions of each nitrate are mixed in the desired proportion, then a half
volume of ethanol is added. Oxalic acid diluted with ethanol is dripped slowly into the nitric solu-
tion, leading to the precipitation of the PLZT oxalate. Thermal dissolution is carried out at 800°C.
In the above coprecipitation methods, thermal dissolution of the precipitate is finally required
to obtain the desired powder. However, the oxide powder can be obtained directly through the
precipitation process in special cases. A direct precipitation method for BaTiO3 is introduced here.
By dripping Ti(OR)4 (R: propyl) into Ba(OH)2 water solution, fine high-purity stoichiometric BT
powders with an average particle size of 10 nm can be obtained directly.
3.2.1.3 Alkoxide Hydrolysis

When metal alkoxides M(OR)n (M: metal atom, R: alkyl) are mixed in alcohol in an appropriate
proportion, and water is added, the hydrolytic reaction produces alcohol and the metal oxide or
metal hydrate. This is sometimes called the sol-gel method. The alkoxide method can produce very
fine and very high purity powders. Since metal alkoxides are evaporative, purification is easily
accomplished through distillation. Also high purity can be sustained during the hydrolic reaction
because no other ion doping is necessary. The mechanisms of hydrolysis and condensation can be
summarized as follows:
(a) Hydrolysis
H − O + M − OR → H − O − M + ROH
|
H

(b) Alkoxylation (removal of H as an alcohol)
M − O + M − OR → M − O − M + ROH
|
H
(c) Oxolation (removal of H as water)
M − O + M − OH → M − O − M + H 2O
|
H
In order to obtain BT powders, Ba(OC3H7)2 and Ti(OC5H11)4 are diluted with isopropyl alcohol (or
benzene). Choosing an appropriate hydrolysis condition (pH), very fine stoichiometric BT powders
with good crystallinity and a particle size of 10–100 Å can be obtained (agglomerate size = 1 μm).
Also the hydrolytic process gives a specimen powder with purity more than 99.98%, leading to a
remarkable increase in the permittivity of the sintered ceramic in comparison with samples pre-
pared by oxide-mixing.14 Table 3.3 lists examples of ferroelectric ceramic powders synthesized by
alkoxide hydrolysis.
In the case of PZT preparation, Pb alkoxide is relatively difficult to obtain compared to Ti and
Zr alkoxides. Thus, one way is to employ a two-stage method for convenience: (Zr,Ti)O2 is prepared
by the alkoxide method, and PbO is added later for use in the solid-state reaction.15 A partial sol-gel
method combined with inexpensive readymade nanosize powders will be a promising direction to
achieve reduction in production costs. Of course, the other way is to use Pb alkoxide; zirconium
n-butoxide Zr(O(CH2)3CH3]4 and titanium isopropoxide Ti[OCH(CH3)2]4 are added into lead acetyl
acetonate Pb(CH3COCHCOCH3) to obtain PZT precursor.
3.2.2 SINTERING PROCESS

After being shaped into a desired shape, the agglomerated powder body is fired at a high tempera-
ture (less than the melting temperature). Accelerated diffusion of the constituent atoms on the fine
particle surfaces due to the surface energy (surface tension) promotes crystal bonding at the contact
interface between the two adjacent particles and provides sufficient mechanical strength to the
ceramic without significant distortion from the initial molded shape. This firing process is called
“sintering,” which primarily eliminates pores and increases the ceramic density (see Figure 3.16).
Notice that the physical properties of the sintered body depend not only on the property of each
fine crystalline particle, but also on the grain boundary and the pores. An example is found in the
mechanical strength: mechanical fracture in ceramic bodies occasionally occurs at the grain bound-
ary (intergranular type). On the contrary, when the crystal itself has a strong cleavage character, the
polycrystalline material shows higher mechanical strength.
TABLE 3.3
Ferroelectric Powders Synthesized by
Hydrolysis of Alkoxides
Crystalline BaTiO3
Ba(Zr,Ti)O3
(Ba,Sr)TiO3
Amorphous Pb(Mg1/3Nb2/3)O3
Ba(Zn1/3Nb2/3)O3
Pb(Zr,Ti)O3
(Pb,La)(Zr,Ti)O3

Diffusion
Neck
Raw powder Grain Grain boundry
Pore
Sintering
Molded body Sintered body
FIGURE 3.16 Schematic diagram of sintering process.
During sintering, the grains grow and the grain shape also changes significantly. However, it is
well recognized that the raw powder characteristics strongly affect the manufacturing conditions
and the final product characteristics. In general, the sintering is accelerated with decreasing par-
ticle size of the raw powder (i.e., with increased specific surface area), because the driving force of
sintering is related to the surface energy of the particles. Moreover, for fine powders, the necessary
diffusion length of the atoms for sintering becomes shorter, which accelerates pore diffusion. This
results in high-density ceramics.
There have been many studies on grain growth. Reference [16] is recommended for further study.
The following relationship between the grain size D and the sintering period t is generally used:
Dβ − D0β ∝ t. (3.3)
In the case of normal grain growth, β = 2, and for abnormal grain growth, β = 3. Figure 3.17 shows
the microphotographs of a PLZT 9/65/35 surface sintered at 1200°C for 1 and 16 h, starting from the
oxalic acid/ethanol method.17 Figure 3.18 shows a good linear relation between the sintering period
and the square of the grain size.
Doping must also be considered. The main role of dopants is to decrease the sintering tem-
perature, but additional effects such as suppression or enhancement of grain growth are sometimes
observed. Excess PbO or Bi2O3 in Pb-based perovskites (PZT) inhibits grain growth. Dy doping
(0.8 at.%) is very effective in suppressing the grain size to less than 1 μm in BaTiO3.18
EXAMPLE PROBLEM 3.1

Ions K1+, Bi3+, Zn2+, and Nb5+ make a disordered complex perovskite crystal, the former two
occupying the A-site and the latter two occupying the B-site. Determine the possible composi-
tion formulas.
Solution
Taking into account the charge neutrality of ABO3 = (K1+ 3+ 2+ 5+ 2–
1− x Bi x )(Zn1− y Nb y )O3 :
+1 × (1 − x ) + 3 × x + 2 × (1 − y) + 5 × y = +6. (P3.1.1)

5μm
(a)
5 μm
(b)
FIGURE 3.17 Grain growth in the PLZT ceramics 9/65/35 ceramics sintered for (a) 1 h and (b) 16 h.
D2 – 0.601 = 0.62
(Grain size D2) (μm2)
10
0 4 8 12 16 20
Sintering time (h)
FIGURE 3.18 Grain growth in PLZT as a function of sintering time.

Then
2 x + 3 y = 3 (0 < x < 1/3 < y < 1). (P3.1.2)
We obtain continuous solutions for x and y, which satisfy Equation P3.1.2. A formula (K3/4 Bi1/4)
(Zn1/6 Nb5/6)O3 is one of the solutions.
3.2.3 SINGLE CRYSTAL GROWTH

3.2.3.1 Quartz, LN, and LT
The popular single crystals are quartz, grown by hydrothermal synthesis, and LiNbO3 and LiTaO3,
by the Czochralski method. Nakamura et al. reported on a monomorph bending actuator fabricated
from a thin plate of LiNbO3 crystal, in which half of the thickness is reverse polarized to function
like a bimorph.19 Although this device is fragile and the bending displacement is not large, linear-
ity in the displacement curve without hysteresis is attractive for some special applications such as
scanning tunneling microscopes.
3.2.3.2 PZN–PT, PMN–PT, and PZT

Single crystal growth of PZT was rather difficult with a composition near the MPB. Thus,
Pb(Zn1/3Nb2/3)O3–PbTiO3 (PZN–PT) and Pb(Mg1/3Nb2/3)O3–PbTiO3 (PMN–PT) have been the
focus of medical acoustic transducer applications, because large single crystals more than 1 cm3 can
be easily grown by a simple flux method, and enormously high electromechanical coupling factors
(95%) and piezoelectric d constants (1570 × 10 −12 C/N) can be obtained when they are poled in a
special crystal direction.20,21 More recent result is covered in Section 2.4.5.
Without using rather expensive Czochralski, hydrothermal or flux method, inexpensive “tem-
plate grain growth” methods have been introduced for preparing BT, PZN–PT, PMN–PT, PZT, and
Pb-free piezoelectric crystals. Figure 3.19 shows a large PMN–PT single crystal prepared from a
seed BaTiO3 crystal with solid-state single crystal growth (SSCG) [Ceracomp, Korea].22
3.2.4 TEMPLATED GRAIN GROWTH

Crystallographic texturing of polycrystalline ferroelectric ceramics exhibits significant enhance-
ments in the piezoelectric response. Templated grain growth (TGG) is a way for textured ceramics
with single crystal-like properties. In TGG, nucleation and growth of the desired crystal on aligned
single crystal template particles with heating results in an oriented material. To facilitate align-
ment during forming, template particles are anisometric in shape. To serve as the preferred sites
Ceramics
Fully dense or PMN–PT

hot-pressed ceramics Single crystal
“ +”
BaTiO3 seed crystal
PMN–PT ceramics φ 40 × t8 mm
Side view
FIGURE 3.19 Growth of PMN–PT single crystals using SSCG technique. (Courtesy of Ho-Yong Lee, http://
www.ceracomp.com. With permission.)

1000
2000 Oriented LF4
800
Smax/Emax (pm/V)
1500 Oriented LF4
Strain (10–6)
600
1000 PZT–D Unoriented LF4
400
500
200
Unoriented LF4
0 0
0 500 1000 1500 2000 0 50 100 150 200
(a) Electric field (V/mm) (b) Temperature (°C)
FIGURE 3.20 Strain vs. electric field curves (a), and the displacement change with temperature (b), for
textured and randomly oriented piezoceramics, (K,Na,Li)(Nb,Ta,Sb)O3. Note the significant enhancement in
the strain level in the textured ceramic, which reaches almost comparable to the PZT ceramic strain. (From
Saito, Y. et al., Nature, 432, 84, 2004. With permission.)
for epitaxy and subsequent oriented growth of the matrix, the template particles need to be single
crystal and chemically stable up to the growth temperature. Messing et al. reported that the resulting
ceramics show texture levels up to 90%, and significant enhancements in the piezoelectric proper-
ties, in comparison with randomly oriented ceramics.23 For example, the piezoelectric d coefficient
of textured PMN–32.5PT piezoelectrics was about 1150 pC/N, 2 ~ 3 times higher than randomly
oriented ceramics, and as high as 90% of the single crystal values.
Toyota Central Research Lab is using reactive-templated grain growth (RTGG) as a key pro-
cessing technique for preparing textured Pb-free piezoelectric ceramics. Bismuth layer-structured
ferroelectric ABi4Ti4O15 (A = Na0.5Bi0.5, Ca, Sr) ceramics with a highly preferred [001] orientation
were fabricated with enhanced piezoelectric properties using Bi4Ti3O12 platelets. Textured simple
perovskite-type ceramics were also prepared in Bi0.5Na0.5TiO3 (BNT), BaTiO3, and K0.5Na0.5NbO3
(KNN)-based compositions with Lotgering’s factor higher than 0.8, which exhibited enhanced
electromechanical coupling coefficients and piezoelectric constants when compared with their ran-
domly oriented counterparts.24 Figure 3.20a and b shows the strain vs. electric field curves and the
displacement change with temperature for textured and randomly oriented piezoceramics, (K,Na,Li)
(Nb,Ta,Sb)O3.25 Note that significant enhancement in the strain level in the textured ceramic, which
reaches almost comparable to the PZT ceramic strain.
3.3 DEVICE DESIGNING

We will introduce several device designs in this section, including single disks, multilayers, com-
posites, and thin/thick films.
3.3.1 SINGLE DISKS

Single disk devices are not practically found these days because of low efficiency in performance/
space. However, they are still important for the laboratory experiments.
EXAMPLE PROBLEM 3.2

The apparent dielectric constant of a BT-based ceramic was measured to be 500, using a 1 mm
thick disk sample. However, it was found that a thin air gap of 0.5 μm developed between the
sample ceramic and the coated electrodes over most of the electrode area on both sides, because
of lack of skill in fabrication. Estimate the actual dielectric constant.

Solution
The above unskillful mistake is occasionally found for a beginner engineer when alcohol is used
to wipe the ceramic disk after polishing, and it is not dried completely on a hot plate (above
100°C). The engineer should be careful not to make an air gap (even submicron) during the elec-
troding process. Anyhow, let us evaluate the property from this result.
The total capacitance of a ceramic capacitor and two air gap capacitors (both top and bottom
surfaces) connected in series was measured experimentally. When we denote the capacitor area,
thickness, and the air gap by S, d, and δ, respectively, the total capacitance is given by
1/C = 1/( ε 0 εS/d ) + 2/(ε 0 S / δ ) = (1/ ε 0 S )(d /ε + 2δ). (P3.2.1)
Since the apparent dielectric constant was calculated from
C /(ε 0 S/d ) = 500, (P3.2.2)
the following relation is obtained:
(1/ε) + (2δ /d ) = 1/500. (P3.2.3)
Substituting d = 10 −3 m, δ = 0.5 × 10 −6 m, we obtain the real dielectric constant of ε = 1000.

Even though this air gap is very narrow, it dramatically reduces the apparent permittivity by
½ because of small air permittivity.
3.3.2 MULTILAYERS
To achieve a low driving voltage, miniaturization and hybridization of the devices, ferroelectric
ceramic multilayer structures have been investigated intensively for capacitor, actuator, and electro-
optic applications. Key words for the future trend will be “finer” and “hybridization.” Layers thin-
ner than 3 μm, which are currently used in multilayer capacitors, can also be introduced in actuator
devices technologically instead of the present 60–80 μm thick sheets. A typical actuator with 10 mm
length exhibits 10 μm displacement (strain of 0.1%) with a fundamental resonance frequency around
100 kHz. Nonuniform configurations or heterostructures of the materials, layer thickness, or the
electrode pattern can be adopted for practical devices.
There are two techniques for making multilayered ceramic devices: the cut-and-bond method
and the tape-casting method. The tape-casting method has been widely used for multilayer capaci-
tor fabrication, and requires expensive fabrication facilities and sophisticated techniques, but is
suitable for mass production of more than 100,000 pieces per month.
As shown in Figure 3.21, a multilayer structure is composed of alternate ferroelectric ceramic and
internal electrode layers fabricated by cofiring. An adjacent pair of electrodes composes a unit dis-
placement element, which is connected in parallel by the external electrode up to hundreds of layers.
External
electrode
Polarization direction
Internal electrode
FIGURE 3.21 Structure of a multilayer actuator.

Calcined powder
(Binder mixing, vacuumization)
Slurry
(Slip casting)
Green sheet
(Punching)
(Electrode printing)
(Lamination, press, cutting)
Green chip
(Binder evaporation, sintering)

(External electrode printing)
Multilayer device
FIGURE 3.22 Fabrication process for a multilayer ceramic actuator.
Figure 3.22 shows a flowchart of the manufacturing process of the multilayer ceramic actuators.
Green sheets are prepared in two steps: slip preparation of the ceramic powder and a doctor blade
process. The slip is made by mixing the ceramic powder with solvent, deflocculant, binder, and plas-
ticizer. The slip is cast into a film under a special straight blade, a doctor blade, whose distance above
the carrier determines the film thickness. After drying, the film, called a green sheet, has the elastic
flexibility of synthetic leather. The volume fraction of the ceramic in the polymer matrix at this point
is about 50%. The green sheet is then cut into an appropriate size, and internal electrodes are printed
using Ni (for BT-based capacitors), silver–palladium, platinum, or Cu ink (for PZT-based transducers).
Several tens to hundreds of such layers are then laminated, and pressed using a hot press. After cutting
into small chips, the green bodies are sintered at around 1200°C in a furnace, taking special care to
control binder evaporation around 500°C. The sintered chips are then polished, externally electroded,
lead wires are attached, and finally the chips are coated with a waterproof spray.
EXAMPLE PROBLEM 3.3

For a certain applied voltage V, a multilayer structure can amplify the generative displacement
in a piezoelectric actuator keeping its total length. Larger displacement magnification can be
expected in an electrostrictor than in a piezoelectric. Verify this situation, using simplest assump-
tions: x = dE and x = ME2 for piezoelectric and electrostrictive cases, respectively.
Solution
Assume L, ΔL, and n to be the total thickness, displacement, and the number of layers of the
ceramic actuator, respectively.
(a) Piezoelectric
ΔL = Lx = LdE = Ld[V/( L /n)] = ndV . (P3.3.1)
The generative displacement is amplified in proportion to the number of layers n.

(b) Electrostrictor
ΔL = Lx = LME 2 = LM[V/( L /n)]2 = n 2 ( M / L )V 2 . (P3.3.2)

120
Temperature rise ΔT (°C)

100
80
60
40
20
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Ve/A (volume mm3/area mm2)
FIGURE 3.23 Temperature rise vs. Ve/A (3 kV/mm, 300 Hz), where Ve is the effective volume generating the
heat and A is the surface area dissipating the heat.
In this case the generative displacement is amplified in proportion to the square of the number of
layers n (more effective than in a piezoelectric).
Zheng et al. reported on the heat generation from multilayer piezoelectric ceramic actuators of
various sizes.26 The temperature change was monitored in actuators driven at 3 kV/mm and 300 Hz.
Figure 3.23 shows the saturated temperature plotted as a function of Ve/A, where Ve is the effective
volume (electroded portion of the actuator) and A is the surface area. This linear relation is reason-
able because the volume Ve generates the heat and this heat is dissipated through the area A. Thus,
if you need to suppress the heat, a small Ve/A (flat and cylinder shapes are better than cube and solid
rod structures, respectively) is preferred.
3.3.3 UNIMORPHS/BIMORPHS
Unimorph and bimorph devices are defined by the number of piezoelectric ceramic plates: only one
ceramic plate is bonded onto an elastic shim, or two ceramic plates are bonded together, respec-
tively. The bimorphs are mainly considered here. The bimorph causes bending deformation because
two piezoelectric plates are bonded together and each plate produces extension or contraction under
an electric field. This effect is employed for various speakers. The bimorph induces voltage for a
bending deformation, which can be used for accelerometers. Since the fabrication process is simple
(two ceramic plates are just bonded using a resin) and large magnification of the displacement is
easily obtained, this structure has been widely used. However, the drawbacks include a low response
speed (1 kHz) and low generative force due to the bending mode (tensile stress concentrates on the
top of the piezoplate). A metallic sheet (called the shim) is occasionally sandwiched between the two
piezoceramic plates to increase the reliability, that is, the structure is maintained even if the ceramic
fractures (Figure 3.24). Also notice the tapering of the bimorph, which increases the response fre-
quency by maintaining the tip displacement magnitude.
There have been many reports on equations describing the tip displacement and the resonance
frequency. Summaries are provided here. Figure 3.25 illustrates two bimorph designs without shims.
Two poled piezoceramic plates with t/2 in thickness (i.e., t is the total thickness) and L in length
are bonded with their polarization directions opposite to each other (a) or parallel to each other (b).
According to the configuration, the tip displacement δ under a voltage V is provided as follows when
one end is clamped (cantilever condition):
δ = (3/2)d31 ( L2/t 2 )V , (3.4a)
δ = 3d31 ( L2/t 2 )V . (3.4b)

Elastic shim Piezoceramic plate

+
Polarization
direction
(a) – +
(b)
FIGURE 3.24 Fundamental structure of a piezoelectric bimorph.
(a)
(b)
FIGURE 3.25 Two types of piezoelectric bimorphs: (a) the antiparallel polarization type and (b) the parallel
polarization type.
Notice that this difference comes from the electrode gap difference: t in (a) and t/2 in (b), leading to
the double electric field in (b). For both cases the fundamental resonance frequency is determined
by the total thickness t as27
f = 0.161(t/L2 )(ρs11E )−1/2 . (3.5)
As can be anticipated, the bimorph drive is inevitably accompanied by a rotational motion. To

obtain a perfect parallel motion a special mechanism must be employed. Figure 3.26 shows such
Sensor electrode
Support
Piezo actuator part
FIGURE 3.26 Bimorph structure for a perfectly parallel motion with a position sensing feedback function
(Ampex).

a bimorph structure. A complex bimorph proposed by Ampex has divided electrodes electrically
connected oppositely at the tip and bottom (supporting part) parts so as to compensate the canting
angle at the bottom by the opposite bend at the tip.28 The bimorph also included a sensor function:
the sensor electrode can detect the voltage generated in proportion to the magnitude of bend.
EXAMPLE PROBLEM 3.4

Using a PZT–based ceramic with a piezoelectric constant of d31 = −300 pC/N, design a no-shim
bimorph with a total length of 30 mm (5 mm is used for cantilever clamping) which can produce
a tip displacement of 40 μm with 20 V applied. Calculate the response speed of this bimorph.
Here, the density and the elastic compliance of the ceramic are ρ = 7.9 g/cm3 and s11
E = 16 × 10 −12
2
m /N, respectively.
Solution
Considering a certain low applied voltage, type (b) in Figure 3.25 is preferred to type (a) in order
to obtain a large displacement. Substituting L of Equation 3.4b with 25 mm, we get the piezo-
electric plate thickness:
t = L 3 (d31 V δ )
(
= 25 × 10 −3 m ) 3 ⋅ (300 × 10 −12
) (
C/N ⋅ (20 V ) 40 × 10 −6 m . )
= 530 μm (P3.4.1)
After cutting the ceramic into plates of 265 μm in thickness, 30 mm in length, and 4–6 mm in
width, the two plates are bonded together after electroding and electrical poling. The width of
the bimorph is usually chosen as w/L < 1/5 so as not to suppress the magnitude of bending. The
response time is estimated by the resonance period. From Equation 3.5
( )( )
−1 2
f = 0.161 t L2 ρs11
E
(
= 0.161 ⎡530 × 10 −6 m 25 × 10 −3 m ⎤ ) (7.9 × 10 )( )
2
3
kg/m 3 16 × 10 −12 m 2 /N
⎢⎣ ⎥⎦
= 378 Hz. (P3.4.2)
This is roughly 2.6 ms. This bimorph was designed for the mechanical camera shutter
originally.
EXAMPLE PROBLEM 3.5

When a piezoceramic plate is bonded to a metallic shim, a unimorph bending device can be
fabricated.29 The tip deflection δ of the unimorph supported in a cantilever style is given by
δ = (d31E )L2Yc tc /[Ym {t02 − (t0 − tm )2} + Yc {(t0 + tc )2 − t02 }], (P3.5.1)
where
E is the electric field applied to the piezoelectric ceramic
d31 is the piezoelectric constant
L is the length of this unimorph
Yc or Ym is Young’s modulus for the ceramic or the metal
tc or tm is the thickness of each material
In addition, t0 refers to the distance between the strain-free neutral plane and the bonding sur-
face, which is represented as

t0 = [tc tm2 (3tc + 4tm )Ym + tc4 Yc ] / [6tc tm (tc + tm )Ym ] (P3.5.2)
Suppose Yc = Ym (PZT and steel have a relationship close to this condition), calculate the opti-
mized condition of (tm /tc) to maximize the deflection δ for the following conditions:
(a) For a fixed ceramic thickness tc (this situation happens when you are a customer of the
ceramic, and need to choose from the company catalogue).
(b) For a fixed total thickness (tc + tm) [this situation happens when you are a manufacturer
of the bimorph component (such as speakers), and the customer requests you to design a
certain resonance frequency device].
Solution
Setting Yc = Ym, the equations become
δ = (d31E )L2 tc /[{t02 − (t0 − tm )2} + {(t0 + tc )2 − t02 (P3.5.3)
t0 = [tc tm2 (3tc + 4tm ) + tc4 ] / [6tc tm (tc + tm )]. (P3.5.4)
Substituting t0 in Equation P3.5.3 with Equation P3.5.4
δ = (d31E )L2 3tm tc /tm + tc )3 . (P3.5.5)
Then, the function f(tm) = tmtc/(tm + tc)3 must be maximized for a fixed ceramic thickness tc (a) or
for a fixed total thickness tc + tm = ttot.
(a) df (tm )/dtm = (tc − 2tm )tc /(tm + tc )4 = 0 (P3.5.6)
Thus, the metal plate thickness should be adjusted to tm = tc/2. T0 = tc/2
(b) df (tm )/dtm = (t tot − 2tm )/t tot

3
=0 (P3.5.7)
Both the metal and ceramic plate thickness should be adjusted to tm = tc = ttot/2; t0 = ttot/3.
This problem suggests that the elastic shim thickness should be adjusted around ½ ~ 1 of the
piezoceramic thickness. We can find unreasonable thickness ratio from this sense in microelectro-
mechanical system (MEMS) cantilever designs reported in many previous papers.
How can we call a bending device which has four layers of PZT plates?
Answer
The following definitions are pictured in Figure 3.27:
Monomorph—single actuator ceramic plate
Unimorph—single actuator plate and an elastic shim
Bimorph—double actuator plates bonded together with or without an elastic shim
Multimorph—multiple actuator plates bonded together with or without multiple elastic shims
A device with four layers of PZT plates is called a “multimorph.” The monomorph is classified
into two:
1. Single crystal LiNbO3 plate with a half-reversed-polarization region [see Section 3.2.3]19
2. Semiconductive BaTiO3 ceramic plate with a nonuniform electric field distribution due to
the Schottky barrier30

(a) Monomorph
(b) Unimorph
(c) Bimorph
(d) Multimorph
FIGURE 3.27 Structures of bimorph family.
3.3.4 FLEXTENSION/HINGE-LEVER AMPLIFICATION MECHANISMS

Market research conducted by the author in the late 1990s suggested that the largest market/demand
for the actuators’ specifications will be as follows:
(a) Stroke = 100 μm

(b) Force = 100 N
(c) Response time = 100 μs
Neither multilayer nor bimorph actuators can satisfy the above specifications directly. A multilayer
needs a displacement amplification mechanism, while a bimorph is required to improve its force
and response speed.
Hinge-lever and flextension-type displacement amplification mechanisms are adopted for the
multilayer actuators. The most famous hinge-type design is the one utilized in a dot-matrix printer
developed by NEC, as depicted in Figure 3.28a (see Section 7.6).31 Figure 3.28b shows flextensional
type developed by Cedrat.32
A composite actuator structure called the “moonie or cymbal” has been developed to amplify the
pressure sensitivity and the small displacements induced in a piezoelectric ceramic.33 The moonie/
cymbal has intermediate characteristics between the conventional multilayer and bimorph actua-
tors; it exhibits an order of magnitude larger displacement (100 μm) than the multilayer, and much
larger generative force (10 kgf) with quicker response (100 μs) than the bimorph. The moonie device
consists of a thin multilayer ceramic element and two metal plates with a narrow moon-shaped
Platen
Paper
Ink ribbon
Guide
Piezoelectric
actuator
Wire
(a) Stroke amplifier (b)
FIGURE 3.28 (a) A hinge lever-type displacement amplification mechanism (NEC) and (b) a flextensional
amplified piezoelectric actuator (Cedrat).

PZT
(a) Metal endcap (b)
FIGURE 3.29 Structures of a moonie (a) and a modified design cymbal (b).
cavity bonded together (Figure 3.29a). The moonie with a size of 5 mm × 5 mm × 2.5 mm can gen-
erate a 20 μm displacement under 60 V, which is 8 times as large as the generative displacement of
a multilayer of the same size. By modifying the end cap design (cymbal type) as shown in Figure
3.29b, twice the displacement can be obtained.34 Also the generative displacement is rather uniform,
independent of the position from the center of the end cap. Another advantage of the cymbal to the
moonie is its easy fabrication process. One-step punching can make end caps from a metal plate.35
3.3.5 FLEXIBLE COMPOSITES

When needle- or plate-shaped piezoelectric ceramic bodies are arranged and embedded in a
polymer matrix, functional composites can be fabricated, which provide enhanced sensitivity by
keeping the actuation function. Figure 3.30a shows such a 1-3 composite device, where PZT rods are
arranged in a polymer in a two-dimensional (2-D) array.
The simplest composite from a fabrication viewpoint is a 0-3 connectivity type, which is made by
dispersing piezoelectric ceramic powders uniformly in a polymer matrix (Figure 3.30b). The fabri-
cation processes are classified into melting and rolling methods.36 Figure 3.31 shows a flowchart for
the fabrication processes. The powders are mixed with molten polymer in the first method, while the
powders are rolled into a polymer using a hot roller in the second method. The connectivity concept,
fabrication processes, and the performance for 1-3 composites are detailed in Section 10.3.
3.3.6 THIN/THICK FILMS

3.3.6.1 Film Manufacturing Techniques
Techniques for fabrication of oxide thin films are classified into physical and chemical processes:
(a) Physical processes

Electron beam evaporation
RF sputtering, DC sputtering
Ion beam sputtering
Ion plating
Piezoceramic fiber (phase 1) Piezoceramic powder (phase 1)
(a) Polymer matrix (b)

(phase 2)
FIGURE 3.30 PZT:polymer composites: (a) 1-3 connectivity and (b) 0-3 connectivity.

Melting method Rolling method
Melting polymer Ferroelectric powder Polymer Ferroelectric powder
(Ball milling) (Rolling)
(Film casting) (Calendering)
Composite film Composite sheet
Electroding
Poling
Piezoelectric component
FIGURE 3.31 Fabrication process for PZT:polymer composites.
(b) Chemical processes

Sol-gel method (dipping, spin coating, etc.)
Chemical vapor deposition (CVD)
Metal-oxide CVD (MOCVD)
Liquid-phase epitaxy, melting epitaxy, capillary epitaxy, etc.
Sputtering has been most commonly used for ferroelectric thin films such as LiNbO3, PLZT,37 and
PbTiO3.38 Figure 3.32 shows the principle of a magnetron sputtering apparatus. Heavy Ar plasma
ions bombard the cathode (target) and eject its atoms. These atoms are deposited uniformly on
the substrate in an evacuated enclosure. Choosing a suitable substrate and deposition condition,
Heater
Ar
Holder
Substrate
Gas Vacuum pump
O2 Magnetic
Plasma
field
Target
N S N
S N S
High-f power supply
FIGURE 3.32 Principle of a magnetron sputtering apparatus.

800
PZT(100) PZT(111) PZT(200)

600
CPS (arbitrary units)

Temperature (°C)
400 PZT(100)
PZT(111)
200
0
0 20 40 60 80 100 20 25 30 35 40 45 50
(a) Annealing time (s) (b) Angle 2θ
FIGURE 3.33 Epitaxially grown rhombohedral (70/30) PZT films with (001) and (111) orientations: (a) opti-
mum rapid thermal annealing profiles, and (b) x-ray diffraction patterns for films grown according to these
profiles. (From Kalpat, S. et al., Jpn. J. Appl. Phys., 40, 158, 2001.)
single crystal-like epitaxially deposited films can be obtained. The sol-gel technique has also been
employed for processing PZT films.39 Applications of thin film ferroelectrics include memories,
surface acoustic wave devices, piezosensors and micromechatronic, or MEMS devices.
As we discussed in Section 2.4.5, (001) epitaxially oriented PZT rhombohedral composition
films are most suitable from the application viewpoint. Kalpat et al. demonstrated (001)- and (111)-
oriented films on the same Pt-coated Si substrate by changing the rapid thermal annealing profile.40
Figure 3.33a and b shows the PZT (70/30) films with (001) and (111) orientations.
3.3.6.2 MEMS Applications

The micromachining process used to fabricate the PZT micropump is illustrated in Figure 3.34.40
The etching process for the silicon:PZT unit is shown on the left-hand side of the figure and that
for the glass plate on the right side. A schematic of the micropump for blood tester is pictured in
Figure 3.35. The blood sample and test chemicals enter the system through the two inlets, identified
in Figure 3.35, are mixed in the central cavity, and finally are passed through the outlet for analy-
sis. The movement of the liquids through the system occurs through the bulk bending of the PZT
diaphragm in response to the drive potential provided by the interdigital surface electrodes.40
Pt/Ti/silicon on
insulator wafer (SOI)
Bottom glass plate
PZT thin film sputtering
Top electrode Au/Ti deposition Masking and wet etching

formation of cavity in glass
Top electrode patterning
Anodic bonding to silicon
(photolithography and lift-off)
wafer
Deep reactive ion-etching (DRIE)

membrane formation
FIGURE 3.34 The micromachining process used to fabricate a PZT micropump. (From Kalpat, S. et al.,
Jpn. J. Appl. Phys., 40, 158, 2001.)

Top Bottom
electrode PZT IDTs electrode
SiO2/Si
Inlet Inlet
Glass
wafer Outlet
FIGURE 3.35 A schematic diagram of the structure of a PZT micropump. Actual size: 4.5 mm × 4.5 mm ×
2 mm. (From Kalpat, S. et al., Jpn. J. Appl. Phys., 40, 158, 2001.)
3.3.6.3 Constraints in Thin/Thick Films

The thin film structure is inevitably affected by four significant parameters:
1. Size constraints: Similar to a powder sample, there may exist a critical film thickness
below which the ferroelectricity would disappear. No research has been made from this
viewpoint yet. Refer to Section 3.4.
2. Stress from the substrate: Tensile or compressive stress is generated due to thermal expan-
sion mismatch between the film and the substrate, leading to sometimes a higher coercive
field for domain reorientation. Curie temperature is also modified with a rate 50°C per
1 GPa.
3. Epitaxial growth: Crystal orientation dependence should be also considered, similar to
the case in single crystals. An example can be found in a rhombohedral composition PZT,
which is supposed to exhibit the maximum performance when the Ps direction is arranged
57° cant from the film normal direction [i.e., (001) crystallographic orientation].
4. Preparation constraint: Si substrate requires low sintering temperature of the PZT film.
Typically 800°C for a short period is the maximum for preparing the PZT, which may limit
the crystallization of the film, leading to the reduction of the properties. A metal electrode
on a Si wafer such as Pt also limits the crystallinity of the PZT film.
3.4 SIZE EFFECT ON FERROELECTRICITY

So-called “fine ceramic” used in 1980s has been rephrased as “nanotechnology” in 2000s; that is,
manufacturing ceramics which possess controlled-grain size or very fine grains (note that “nano”
does not mean a real nanometer range, but a micron range). In parallel, so-called “amorphous fer-
roelectric” was studied in 1980s. Lines theoretically suggested that significantly large permittivity
might be realized when a ferroelectric ceramic is prepared in an amorphous form.41 Note that it
did not suggest that amorphous form of this material exhibits “ferroelectric” properties. All results
reported during the 1980s period concluded that there is little possibility to realize “amorphous”
ferroelectrics. Different from “ferromagnetics” which originates from short-range spin-exchange
coupling, “ferroelectrics” originates from a cooperative phenomenon based on rather long-range
Coulombic coupling. In conclusion, relatively large crystalline size (submicron meter) seems to
be required to realize “ferroelectricity,” by annealing amorphous ferroelectrics to increase the
crystallinity.42
Small particle “nano” technology seems to eliminate the ferroelectric functionality, unfortu-
nately. However, thin film technology with nanometer thickness seems to be liable to realize good

ferroelectricity. There remains a future question to clarify the following from both experimental and
theoretical viewpoints:
1. 3-D problem: What is a critical particle size below which ferroelectricity will disappear?
2. 2-D problem (1-D connectivity): What is a critical fiber diameter below which ferroelec-
tricity will disappear?
3. 1-D problem (2-D connectivity): What is a critical film thickness below which ferroelectricity/
piezoelectricity will disappear?
3.4.1 GRAIN SIZE EFFECT ON FERROELECTRICITY

To understand the grain size dependence of the dielectric properties, we must consider two size
regions: The micrometer range in which a multiple domain state becomes a monodomain state, and
submicrometer range in which the ferroelectricity becomes destabilized.
Figure 3.36 shows the transverse field-induced strains of 0.8 at.% Dy-doped fine grain ceramic
BaTiO3 (grain diameter around 1.5 μm) and of the undoped coarse grain ceramic (50 μm), as
reported by Yamaji.18 As the grains become finer, under the same electric field, the absolute value
of the strain decreases and the hysteresis becomes smaller. This is explained by the increase in
coercive field for 90° domain rotation with decreasing grain size. The grain boundaries (with many
dislocations on the grain boundary) “pin” the domain walls and do not allow them to move easily.
Also the decrease of grain size seems to make the phase transition of the crystal much more diffuse.
Figure 3.37 shows the temperature dependence of the piezoelectric coefficient d33. Although the
absolute value of d33 decreases in the Dy-doped sample, the temperature dependence is remarkably
improved for practical applications. It should be noted that Yamaji’s experiment cannot separate the
effect due to intrinsic grain size from that due to dopants.
1 × 10–4
Electric field (kV/mm)

–1.5 –1.0 –0.5 0.5 1.0 1.5
0.8 at.% Dy-doped
BaTiO3 ceramic
–1
Transverse strain x1
Undoped BaTiO3 ceramic
–2
–3
FIGURE 3.36 Electric-field-induced strain curves in Dy-doped and undoped BaTiO3 ceramic samples.
(From Yamaji, A. et al., Proc. 1st Mtg. Ferroelectric Mater. Appl., Kyoto, 269, 1977.)

200
Undoped
coarse grained BT
d33 (10–12 C/N)

100
Dy-doped fine-grained BT
0
0 50 100
Temperature (°C)
FIGURE 3.37 Temperature dependence of the piezoelectric d33 in Dy-doped and undoped BaTiO3 ceramics.
(From Yamaji, A. et al., Proc. 1st Mtg. Ferroelectr. Mater. Appl., Kyoto, 269, 1977.)
1 kHz
100 kHz
10 Temp rise
Peak permittivity (×103)
Temp fall
0
0 1 2 3 4 5
Grain size (μm)
FIGURE 3.38 Grain size dependence of the peak permittivity in PLZT 9/65/35. (From Uchino, K. and
Takasu, T., Inspection, 10, 29, 1986.)
Takasu et al. studied the effects of grain size on PLZT.43 They obtained PLZT (9/65/35) powders
by coprecipitation. Various grain sizes were prepared by hot-pressing and by changing sintering
periods, without using any dopants. PLZT (9/65/35) shows significant dielectric relaxation (fre-
quency dependence of the permittivity) below the Curie point of about 80°C, and the dielectric
constant tends to be higher at lower frequency. Figure 3.38 shows the dependence of the peak dielec-
tric constant on grain size. For grain size larger than 1.7 μm, the dielectric constant decreases with
decreasing grain size. Below 1.7 μm, the dielectric constant increases rapidly. Figure 3.39 shows
the dependence of the longitudinal field-induced strain on the grain size. As the grain size becomes
smaller, the maximum strain decreases monotonically. However, when the grain size becomes less
than 1.7 μm, the hysteresis is reduced. This behavior can be explained as follows: with decreas-
ing grain size, (anti)ferroelectric (ferroelastic) domain walls become difficult to form in the grain,
and the domain rotation contribution to the strain becomes smaller (multidomain–monodomain

Grain size D = 4.5 μm

20
D = 2.4 μm
16
Longitudinal strain x3 (×10–4)

D = 1.1 μm
12
–15 –7.5 0 7.5 15

Applied field (kV/cm)
FIGURE 3.39 Grain size dependence of the induced strain in PLZT ceramics. (From Uchino, K. and Takasu,
T., Inspection, 10, 29, 1986.)
transition model). The critical size is about 1.7 μm. However, note that the domain size is not con-
stant, but is dependent on the grain size, and that in general the domain size decreases with decreas-
ing grain size.
3.4.2 3-D PARTICLE SIZE EFFECT ON FERROELECTRICITY

Regarding the much smaller grain/particle size range, Uchino et al. reported previously a number of
informative experiments. Figure 3.40 shows the most cited figure in recent nanotechnology papers.
The degree of tetragonality (i.e., c/a ratio) is plotted as a function of particle size in pure BaTiO3 at
room temperature.44 The c/a value decreases drastically below 0.2 μm and becomes 1 (i.e., cubic)
at 0.12 μm, defined as a critical particle size. Figure 3.41 shows the temperature dependence of the
c/a ratio for various particle size powders. This demonstrates the correlation between the critical
Single crystal
1.010
1.008
Tetragonality (c/a)
1.006
1.004
1.002
1.000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Particle size (μm)
FIGURE 3.40 Particle size dependence of the tetragonality in BaTiO3 at room temperature. Note that the
tetragonal crystal distortion disappears below a critical particle size around 0.1 μm. (From Uchino, K. et al.,
J. Am. Ceram. Soc., 72, 1555, 1989.)

1.0100
S-1
C1-30.3 μm
1.0 μm
C-2
0.22 μm
1.0075
Tetragonality (c/a) C-1 0.18 μm
H-4 0.16 μm
1.0050
H-3 0.14 μm
H-2 0.12 μm
1.0025
1.0000
0 50 100 150
Temperature (°C)
FIGURE 3.41 Temperature dependence of the tetragonality in BaTiO3 for various particle size samples.
(From Uchino, K. et al., J. Am. Ceram. Soc., 72, 1555, 1989.)
particle size and the Curie temperature, which decreases with decreasing particle size. This chapter
clearly indicated that the ferroelectricity disappears on reducing the 3-D particle size. It is interest-
ing to note that the tetragonality between 1 and 1.0025 was not observed experimentally, suggesting
that the first-order phase transition seems to be sustained regardless of the crystalline size. The
tetragonality change curve is simply shifted to lower temperature when reducing the particle size.
Though multiple recent papers discussed on the critical particle size of 0.12 μm, too small or
too large, experimentally, it is not the keypoint, because there is a big ambiguity in determining
the average particle size as Ref. [45] explicitly explained. Specific surface area, x-ray diffraction,
electron microscope, or optical microscope provides large deviation on the particle size more than
several times.
A similar critical particle size was also reported in (Ba,Sr)TiO3,45 (Ba,Pb)TiO3,46 and for antifer-
roelectric PbZrO3.45 Figure 3.42 shows the relationship between the critical particle size Dcrit and the
Curie (or Neel) temperature TC for these materials. An important empirical rule was obtained:
Dcrit × (TC − Room temp.) = Constant. (3.6)
Although there have been various reports on the critical grain/particle size, it seems to be at least true that
there is a kind of critical size below which the ferroelectricity disappears (i.e., the crystal becomes cubic).
No satisfactory explanation has yet been presented so far. One possible explanation is based on
a hydrostatic pressure model. In general, the ferroelectric transition temperature decreases sharply
(50°C/MPa) with increasing hydrostatic pressure. Systematic data can be found in Samara’s papers,
who used external hydrostatic oil pressure.47 The effective surface tension γ on a fine particle causes
a hydrostatic pressure p intrinsically48:
p = 2 γ/R ( R :particle radius). (3.7)

Critical particle size 2Rcrit (μm)

Hydrostatic pressure P (GPa)
10 0.2
5 0.1
PbZrO3
0 0
0 100 200 300 400 500
Curie/néel temperature (°C)
FIGURE 3.42 Relationship between the critical particle size Dcrit or the critical hydrostatic pressure and the
phase transition temperature.
TABLE 3.4
Critical Particle Size, Critical Hydrostatic Pressure, and the Surface
Tension Energy for Various Perovskites
Curie
Material Temperature (°C) 2Rcrit (mm) Pcrit (GPa) g (N/m)
Ba0.9Sr0.1TiO3 95 0.19 1.2 57
BaTiO3 125 0.12 1.8 54
Ba0.85Pb0.15TiO3 180 0.08 2.9 58
Ba0.5Pb0.5TiO3 330 0.032 6.2 50
PbTiO3 500 0.02 10 50
Note: P = 2γ/R.
From the critical particle size Dcrit (=2Rcrit) and the critical hydrostatic pressure pcrit, above which
the cubic structure is realized at room temperature, we can calculate the effective surface tension
γ as listed in Table 3.4. The γ value is almost constant for all perovskite ferroelectrics. The γ value,
50 times larger than that of nonpolar oxides (such as MgO), may be due to the additional energy
from the surface charge contribution, and/or from a crystallographically different skin phase on the
particle surface (i.e., core-shell model).
Why cannot we expect amorphous ferroelectricity or ferroelectric nanoparticles?

Answer
To understand the reason why the ferroelectricity will disappear with decreasing particle size,
we can consider the energy fluctuation for a nanosize ferroelectric particle as follows (Refer to
Problem 1.1). Consider a 1-D finite chain of two kinds of ions +q and −q, arranged alternately with
a distance a (see Figure 3.43a). A nanosize crystal grows gradually, starting from a single positive
ion, and adding a pair of negative or positive ions, thus keeping a crystal size 2na (n = 1, 2, 3, …).
With increasing crystal size, the crystal Coulomb energy will be changed continuously as
U1 = (2/4πε0ε)[−(q 2/a)]
U2 = (2/4πε0ε)[−(q 2/a) + (q 2/ 2a)]

–q +q –q +q –q +q –q
–3a –2a –a 0 +a +2a +3a

(a)
1 3 5 7 9 n
Madelung energy
Energy U
–(2/4πε0)(q2/a)
(b)
FIGURE 3.43 1-D finite chain of two kinds of ions +q and −q.
U3 = (2/4πε 0 ε)[ −(q 2/a) + (q 2/a) + (q 2/2a) − (q 2/3a)]
...
Un = (2/4πε 0 ε)[ −(q 2/a) + (q 2/2a) − ... + (q 2/na)].
The Coulomb potential change at the center point is shown in Figure 3.43b as a function of the
crystal size n, with a final saturating value of Madelung energy. When the energy fluctuation is so
large, a small energy imbalance between a paraelectric and a ferroelectric state may not cause a
phase transition. Supposing that the minimum crystal size (2na) is required to maintain the poten-
tial energy fluctuation less than ±10% at the center positive ion, even when adding or subtracting
a pair of ions (i.e., n + 1 or n − 1). If the basic crystal Coulomb energy is not stabilized to less
than this degree of fluctuation, we cannot expect the ferroelectric phase transition to occur as a
cooperative phenomenon.
By equating 1/n = 10%, we get n = 10. If we use a = 4 Å, 2na = 80 Å = 8 nm. According to
the study on amorphous PbTiO3,49 the soft phonon mode and the maximum permittivity which
indicate the ferroelectricity appearance started to be observed around 100 Å, on increasing the
crystalline size during annealing process. This crystalline size is in the same range as the above
theoretically estimated crystal size. The reader is requested to extend the theory to 3-D model.
Chapter Essentials
1. The MPB composition is the key to enhance the ferroelectric and piezoelectric properties.
2. Doping effects on ferroelectricity in PZT:
Acceptor ion → domain pinning → “hard” piezoelectric
Donor ion → Pb deficiency compensation → “soft” piezoelectric
3. Preparation of ceramic powders:
Oxide-mixing technique
Coprecipitation
Alkoxide hydrolysis
4. Device designs:
Single disk
Multilayer

Unimorph/bimorph
Moonie/cymbal
Flexible composite
Thin/thick film
5. Comparison between multilayers and bimorphs:
1. The multilayer type does not exhibit large displacements, but has advantages in gen-
erative force, response speed, lifetime, and electromechanical coupling k33.
2. The bimorph type exhibits large displacements, but shows disadvantages in generative
force, response speed, lifetime, and the electromechanical coupling keff.
3. Neither multilayer nor bimorph satisfies directly the current demand for the specifica-
tions (displacement: 100 μm, force: 100 N, response: 100 μs).
6. Tip displacement δ in a bimorph under a one-end clamp condition (cantilever):
δ = (3/2)d31 (L2 /t 2 )V or δ = 3d31(L2 /t 2 ) V , according to the structure.
Fundamental resonance frequency is determined by the length and total thickness t:
f = 0.16(t /L2 )(ρs11E )−1/2
7. Particle size dependence of ferroelectricity:
There seems to be a critical particle/grain size (submicrometer), below which the material
ceases to be ferroelectric (i.e., it becomes paraelectric).
Smaller grain → smaller polarization and strain, smaller hysteresis → larger mechanical
strength
Check Point
1. What is the MPB composition of the PZT system at room temperature? PbZrO3:PbTiO3 =
(48:52), (50:50), (52:48), or (none of these)?
2. (T/F) The MPB composition of the PZT system exhibits the maximum electromechanical
coupling k, piezoelectric coefficient d, and the minimum permittivity ε. True or False?
3. How to make PZT ceramic “soft.” Nominate a suitable doping ion species.
4. (T/F) When Fe3+ ion is doped in PZT, oxygen deficiencies are created, resulting in “soft”
characteristics of the PZT. True or False?
5. How large displacement can we expect under a reasonably high electric field for a 10 mm
long multilayer PZT actuator? 0.1, 1, 10, or 100 μm?
6. What is the fundamental resonance frequency for a 10 mm long multilayer PZT actuator?
1, 10, 100, or 1000 kHz?
7. (T/F) There are two bimorph actuators with the same PZT composition: Type I: 40 mm
long, 1 mm thick, 6 mm wide, and Type II: 20 mm long, 0.5 mm thick, 3 mm wide. The
resonance frequencies for these two bimorphs are the same. True or False?
8. (T/F) There are two bimorph actuators with the same PZT composition: Type I: 40 mm
long, 1 mm thick, 6 mm wide, and Type II: 20 mm long, 0.5 mm thick, 3 mm wide. When
the same voltage is applied, the tip displacement under a cantilever support for these two
bimorphs are the same. True or False?
9. (T/F) On reducing the grain size of the BaTiO3 ceramic around 1 μm range, the piezoelec-
tric d constant is decreased. True or False?
10. What is the remaining item of four constraints of the ferroelectric films: (1) grain size,
thickness dependence of the properties, (2) epitaxial growth, crystallographic orientation
dependence, (3) preparation constraints such as annealing temperature, and (4) what?

100 μm
100 μm
FIGURE 3.44 PZT multilayer actuators with their layer thicknesses of (a) 280 and (b) 17 μm. (From Schuh,
C., Proc. of New Actuators 2004, Bremen, 2004. With permission.)

Chapter Problems
3.1 Using a PZT-based ceramic plate with a piezoelectric constant, d31 = −300 pC/N,
size, 25 mm × 5 mm × 0.5 mm, and a phosphor bronze plate (high-Qm material),
design a unimorph with a total active length of 25 mm which can produce the maxi-
mum tip displacement under an application of 100 V. Determine the optimized thick-
ness of a phosphor bronze plate, then calculate the maximum displacement. Here,
the density and the elastic compliance of the ceramic are ρ = 7.9 g/cm3 and s11 E =
−12 2
16 × 10 m /N, respectively. Search for the necessary data (Young’s modulus, etc.) of
phosphor bronze. (Refer to Example Problem 3.5.)
3.2 Describe the difference between “silver paste” and “silver ink.”
Hint
Silver paste includes glass particles with silver particles and used for the external elec-
trode. Annealing is required to melt the glass phase to provide the bonding strength.
On the other hand, ink does not include the glass phase, is usually used for the internal
embedded electrode, in order to cover the peeling weakness. Search further information
on the glass composition and the volume percentage.
3.3 Two types of multilayer PZT actuators were prepared with their layer thicknesses of 280
and 17 μm, as shown in Figure 3.44a and b.50
This PZT composition has tetragonal symmetry, and when the x-ray diffraction pat-
tern was taken for both samples from the top surface, Figure 3.45 was obtained. The
sample with 280 μm layer thickness seemed to have no texture by showing the x-ray
intensity ratio roughly 1:2 for the (001) and (100) reflections, while the intensity of the
(001) reflection was very high for the sample with 17 μm layer thickness.
Based on these experimental results, answer to the following questions:
(a) What sort of texture is indicated from the x-ray pattern for the multilayer actuator
with 17 μm thickness?
(b) Why was this texture created by reducing the layer thickness?
(c) Can we expect the improvement of the piezoelectric performance due to this texture?
5000 (002)t
(200)t
4000
d = 17 μm (textured)
d = 280 μm (texture free)
Intensity
3000
2000
1000
0
43.5 44.0 44.5 45.0 45.5 46.0
2θ
FIGURE 3.45 Two x-ray diffraction patterns for the multilayer actuators with 280 and 17 μm layer thick-
nesses. (From Schuh, C., Proc. of New Actuators 2004, Bremen, 2004. With permission.)

d = 280 μm d = 17 μm
FIGURE 3.46 Two domain texture models for the multilayer actuators with 280 and 17 μm layer thicknesses.
(From Schuh, C., Proc. of New Actuators 2004, Bremen, 2004. With permission.)
Tensile stresses Compressive stresses
FIGURE 3.47 Domain texture creation model for the multilayer actuators with 17 μm layer thickness. (From
Schuh, C., Proc. of New Actuators 2004, Bremen, 2004. With permission.)
Hint
The high intensity of the (001) reflection for the sample with 17 μm layer thickness
indicates the polarization direction aligned normal to the layer plane, as schematically
shown in Figure 3.46.
If this domain texture was induced by the mechanical stress through the manufac-
turing process, the thermal expansion difference between the PZT and metal electrode
may be the origin. Because the metal electrode shrinkage is larger than the PZT during
the cooling-down process, the PZT layer will experience compressive stress (the thinner
layer sample will experience larger compressive stress), leading to the domain align-
ment normal to the layer surface (Figure 3.47).
REFERENCES
1. B. Jaffe, W. R. Cook, and H. Jaffe: Piezoelectric Ceramics, p. 142, Academic Press, New York (1971).
2. K. Uchino and S. Nomura: Jpn. J. Appl. Phys., 18, 1493 (1979).
3. K. Abe, O. Furukawa, and H. Inagawa: Ferroelectrics, 87, 55 (1988).
4. A. Hagimura and K. Uchino: Ferroelectrics, 93, 373 (1989).
5. K. Uchino, H. Negishi, and T. Hirose: Jpn. J. Appl. Phys., 28(Suppl. 28-2), 47 (1989).
6. S. Hirose, Y. Yamayoshi, M. Taga, and H. Shimizu: Jpn. J. Appl. Phys., 30(Suppl. 30-1), 1117 (1991).
7. S. Takahashi and S. Hirose: Jpn. J. Appl. Phys., 32(Pt. 1), No. 5B, 2422 (1993).
8. K. Uchino, J. Zheng, A. Joshi, Y. H. Chen, S. Yoshikawa, S. Hirose, S. Takahashi, and J. W. C. de Vries:
J. Electroceram., 2, 33 (1998).
9. S. Hirose, M. Aoyagi, Y. Tomikawa, S. Takahashi, and K. Uchino: Proc. Ultrasonics Int’l. ’95, Edinburgh,
p. 184 (1995).
10. Kato: Fine ceramics technology, in Fabrication Technology of Ceramic Powder and Its Future, vol. 3, p.
166, Industry Research Center, Japan (1983).
11. M. Lejeune and J. P. Boilot: Ferroelectrics, 54, 191 (1984).
12. S. L. Swartz, T. R. Shrout, W. A. Schulze, and L. E. Cross: J. Am. Ceram. Soc., 67, 311 (1984).
13. Tanada, Yamamura, Shirasaki: Abstract 22nd Jpn. Ceram. Soc. Fundam. Div. 3B5, p. 81 (1984).
14. Ozaki: Electron. Ceram., 13, Summer, 26 (1982).

15. Kakegawa, Mohri, Imai, Shirasaki and Tekahashi: Abstract 21st Jpn. Ceram. Soc. Fundam. Div. 2C6,
p. 100 (1983).
16. H. Abe: Recrystallization, Mater. Sci. Series 2, Kyoritsu Pub., Tokyo (1969).
17. K. Uchino and T. Takasu: Inspection, 10, 29 (1986).
18. A. Yamaji, Y. Enomoto, E. Kinoshita, and T. Tanaka: Proc. 1st Mtg. Ferroelectr. Mater. Appl., Kyoto,
p. 269, (1977).
19. K. Nakamura, H. Ando, and H. Shimizu: Jpn. J. Appl. Phys., 26(Suppl. 26-2), 198 (1987).
21. J. Kuwata, K. Uchino, and S. Nomura: Jpn. J. Appl. Phys., 21(9), 1298 (1982).
22. http://www.ceracomp.com/
23. G. Messing, S. Trolier-McKinstry, E. Sabolsky, C. Duran, S. Kwon, B. Brahmaroutu, P. Park, H. Yilmaz,
P. Rehrig, K. Eitel, E. Suvaci, M. Seabaugh, and K. Oh: Crit. Rev. Solid State Mater. Sci., 29(2), 45
(2004).
24. T. Tani and T. Kimura: Adv. Appl. Ceram., 105(1), 55 (2006).
25. Y. Saito, H. Takao, T. Tani, T. Nonoyama, K. Takatori, T. Homma, T. Nagaya, and M. Nakamura, Lead-
free piezoceramics, Nature, 432(4) 84–87 (2004).
26. J. Zheng, S. Takahashi, S. Yoshikawa, K. Uchino, and J. W. C. de Vries: J. Am. Ceram. Soc., 79, 3193
(1996).
27. K. Nagai and T. Konno, eds.: Electromechanical Vibrators and Their Applications, Corona Pub., San
Antonio, TX (1974).
28. K. Uchino: Piezoelectric Actuators and Ultrasonic Motors, p. 241, Kluwer Academic Publishers, Boston,
MA (1997).
29. K. Abe, K. Uchino, and S. Nomura: Jpn. J. Appl. Phys., 21, L408 (1982).
30. K. Uchino et al.: Jpn. J. Appl. Phys., 26, 1046 (1987).
31. T. Yano, E. Sato, I. Fukui and S. Hori: Proc. Int’l Symp. Soc. Inf. Display, p. 180 (1989).
32. http://www.cedrat.com/en/mechatronic-products.html
33. A. Dogan: PhD thesis, Penn State University, University Park, PA (1994).
34. Y. Sugawara, K. Onitsuka, S. Yoshikawa, Q. C. Xu, R. E. Newnham, and K. Uchino: J. Am. Ceram. Soc.,
75, 996 (1992).
35. H. Goto, K. Imanaka, and K. Uchino: Ultrasonic Technol., 5, 48 (1992).
36. Kitayama: Ceramics, 14, 209 (1979).
37. M. Ishida et al.: Appl. Phys. Lett., 31, 433 (1977).
38. M. Okuyama et al.: Ferroelectrics, 33, 235 (1981).
39. S. K. Dey and R. Zuleeg: Ferroelectrics, 108, 37 (1990).
40. S Kalpat, X. Du, I. R. Abothu, A. Akiba, H. Goto, and K. Uchino: Jpn. J. Appl. Phys., 40, 158 (2001).
41. M. E. Lines: Phys. Rev., B15, 497 (1980).
42. T. Nakamura: Solid State Phys., 20(8), 660 (1985).
43. K. Uchino and T. Takasu: Inspec, 10, 29 (1986).
44. K. Uchino, E. Sadanaga, and T. Hirose: J. Am. Ceram. Soc., 72, 1555 (1989).
45. T. Yamakawa and K. Uchino: Proc. Int’l. Symp. Appl. Ferroelectr. ’90, p. 610 (1991).
46. K. Saegusa et al.: Am. Ceram. Soc., 8, 221 (1990).
47. G. A. Samara: Ferroelectrics, 2, 277 (1971).
48. K. Uchino, E. Sadanaga, K. Oonishi, and H. Yamamura: Ceram. Trans. Ceram. Dielect., 8, 107 (1990).
49. M. Takashige and T. Nakamura: Jpn. J. Appl. Phys., 22(Suppl. 22-2), 29 (1983).
50. M. Laurent, H. Bödinger, T. Steinkopff, K. Lubitz, C. Schuh, S. Wagner, M. J. Hoffmann, H. Murmann-
Biesenecker, and A. J. Schmid: Proc. 14th IEEE Int. Symp. Appl. Ferroelectr. ’04, IEEE-UFFC-S,
Montreal, Canada, Aug. 23–27 (2004).

4 High-Permittivity Dielectrics
The major application of ferroelectrics is for capacitors, utilizing their high dielectric constants
around the Curie temperature.
4.1 CERAMIC CAPACITORS

There are two classes of ceramic capacitors: one is for thermal compensation of electric circuits,
using a TiO2-based material, and the other is a high-permittivity capacitor with a BaTiO3- or
Pb(Zr,Ti)O3-based material. More precisely, there are four primary categories of dielectrics:
1. High-Q, low-dielectric constant K (100) temperature-compensating materials (capacitance

change ±30 ppm);
2. Intermediate-K (3000) materials, labeled X7R or BX (±15%);
3. High-K (10,000) formulations, known as Z5U or Z5V (20%–50%);
4. Nonhomogeneous, barrier layer materials that can have effective K up to 100,000.
Figure 4.1 summarizes the various capacitor types, highlighting their sizes and operating frequency
ranges.1 Ceramic capacitors with a single parallel plate design are still popular, while multilayer
ceramic capacitors are 1/30−1/50 the size of the single parallel plate-type. Semiconductor capaci-
tors exhibit large capacitances using very thin dielectric layers in a semiconductor-based ceramic
(see Section 9.3). Microchip capacitors are ultrasmall capacitors for high-frequency applications.
The basic specifications required for capacitors are
(a) Small size, large capacitance

Materials with a large dielectric constant are desired
(b) High-frequency characteristics
Ferroelectrics with a high dielectric constant are sometimes associated with dielectric disper-
sions, which must be taken into account for practical applications
(c) Temperature characteristics
We need to design materials to stabilize the temperature characteristics
EXAMPLE PROBLEM 4.1

Calculate the wavelength in air (ε = 1) and in a dielectric material with ε = 30 for an electromag-
netic wave at 10 GHz.
Solution
Taking account of c = 3.0 × 108 m/s in air and n = c/ 30 in the dielectric,
λ = c f = 3.0 × 108/10 × 109 = 3 × 10 −2 [m] = 3[cm] (in air) (P4.1.1)
= 3 × 10 −2 30 [m] = 5.5[mm] (in dielectric) (P4.1.2)
Notice a drastic shrinkage of the wavelength in a dielectric.
115

Satellite 30 GHz
commun. Micro
Automobile 3 GHz chip
commun. Semi capacitor
UHF TV 300 MHz conductor
Ceramic Multilayer capacitor
VHF TV 30 MHz capacitor ceramic
FM radio capacitor
3 MHz
AM radio
300 kHz
30 kHz
1 cm
FIGURE 4.1 Various capacitor types classified according to their sizes and operating frequency ranges.
4.2 CHIP CAPACITORS

4.2.1 THIN MULTILAYER CAPACITOR
Multilayer structures were invented by Sandia researchers during World War II, aiming at large
charge capacitance for a nuclear bomb switch application (Manhattan Project). The technology
has been further improved recently as part of capacitor manufacturing aimed at the integration of
electrical circuit components. Figure 4.2 schematically shows a multilayer capacitor (MLC) chip.
Thin sheets made by the tape casting technique, starting from a slurry of the dielectric powder and
organic solvents, were previously coated with Ag–Pd, Ag. Now cheaper Ni (or Cu) paste is used to
form the electrodes, and then several tens to hundreds of sheets are stacked together and sintered.
Finally, external electrodes, used to connect the chip with the circuit, are painted on. See Section
3.3.2 for the details of the manufacturing process.
The layer thickness of MLCs has been reduced remarkably, with thickness currently on the order
of 2–3 μm. The electrostatic capacitance of a MLC is given by the following formula:
C = nε 0 εS (T n ) = ε 0 ε (S T )n2 (4.1)
where
ε is the relative permittivity of the dielectric material
n is the number of layers
S is the electrode area
T is the total thickness of the capacitor
Area S External
electrode
Total
thickness
T Electric field
Internal electrode
FIGURE 4.2 Structure of an MLC.

High-Permittivity Dielectrics 117
TABLE 4.1
Dimensions of the MMCs
Dimensions
Capacitance at (mm) Green Sheet
Room Volume Relative Thickness
Temperature (mF) L W T (mm3) Volume (%) (mm)
Present ceramic 1 0.8 0.7 0.4 0.22 (100) 3
capacitor 10 1.6 0.8 0.8 1.02 100 3
Conventional 1 3.3 1.7 1.2 6.7 (3100) 25
ceramic capacitor 10 7.0 4.2 2.4 70.0 6900 30
Tantalum electrolytic 1 3.2 1.6 1.6 8.2 (3700) —
capacitor 10 4.7 2.6 2.1 25.7 2500 —
Note that the capacitance increases in proportion to the square of the number of layers (i.e., n2),
when the total size is fixed. Table 4.1 summarizes the specifications for several MLCs. The con-
ventional capacitor of 10 μF with a 30 μm layer thickness has a volume of 70 mm3. By decreasing
the layer thickness down to 3 μm, the device volume can be reduced to 1 mm3. Note that by reduc-
ing the layer thickness by 1/n, the total volume is reduced by a factor of (1/n)2 to sustain the same
capacitance.
The state-of-the-art for MLC developed by Taiyo Yuden, Japan, in 2005 is shown in Figure 4.3,
which is a Ni-electrode MLC (3216-type, B-type, 100 μF) with 700 layers of 1.1 μm each layer.3
4.2.2 BASE-METAL INTERNAL ELECTRODE

Since BaTiO3-based capacitor ceramics must be sintered at a relatively high temperature, Ag–Pd
internal electrodes were conventionally used. However, the price of Pd increased drastically in
the end of 1990s because of the economic recession of Russia, which is the major producer of Pd
in the world. Therefore, capacitor manufacturers could not continuously use Ag–Pd internal elec-
trodes for the inexpensive MLCs. Thus, they started to use base metals (Ni or Cu) for the internal
electrodes.
However, as shown in Figure 4.4, the Ellingham diagram for Ni and Cu suggested that the base metal
becomes an oxide during a regular sintering condition, and that both a low temperature (>1000°C)
and a reduced atmosphere are required for manufacturing base-metal embedded MLC chips.
4.2.2.1 Barium Titanate–Based Chip

It can be learned from Chapter 9 that barium titanate (BT) ceramics are easily reduced in a reduced
atmosphere, which increases the conductivity of the ceramic, leading to lifetime reduction. How can
we suppress the BT reduction? Codoping of donor and acceptor ions is a possible solution. From this
standpoint, the so-called magic ions, rare earths such as Y, Dy, and Ho, are introduced here.3 Figure 4.5a
shows the lifetime improvement by “magic ion (Dy and Ho)” doping. Figure 4.5b shows a cross-
sectional view of an X7R capacitor with Ni internal electrodes (dielectric constant K ≈ 2800;
10 μF/1206; 300 total layers of 3 μm each).
4.2.2.2 PbTiO3-Based Chip

The situation is different for Pb-contained ceramics and BT-based ceramics; that is, Cu should
be chosen rather than Ni. Figure 4.6a and b compares thermodynamic calculations for PbTiO3 and

10 mm
(a)
1 μm
(b)
FIGURE 4.3 Ni-electrode MLC (3216-type, B-type, 100 μF with 700 layers of 1.1 μm each). (a) Optical cross-
sectional view and (b) SEM image (after thermal etch). (From Kishi, H., Mater. Integr., 19, 47, 2006. With
permission.)
T (°C)
1400 1200 1000 900 800 700 600 500 400
–5 1ppm
log p(O2) Cu(II) CuO–Cu2O

(bar) –10 H2/N2
Ni(II) Cu–Cu2O
Cu(I) Ni–NiO
–15
–20 Ni(m) Cu(m)
–25
–30
0.0006 0.0008 0.0010 0.0012 0.0014 0.0016
–1/T (1/K)
FIGURE 4.4 Ellingham diagram for Ni and Cu.

10,00,000
1,00,000
X7R (10 μm)
10,000 X7R (5 μm)
Life time (s) Y5V (10 μm)
1,000 Y5V (5 μm)
100
10
1
La Nd Sm Gd Dy Ho Er
(a) Rare earth
(b)
FIGURE 4.5 (a) Rare-earth ion (Dy, Ho) doping effect on the lifetimes of MLCs with Ni internal elec-
trode. (b) Cross-sectional view of an X7R capacitor with Ni internal electrodes (dielectric constant K ≈ 2800;
10 μF/1206; 300 total layers of 3 μm each). (From Kishi, H., Mater. Integr., 19, 47, 2006.)
(a) Ni and (b) Cu electrodes. When Ni is used as an electrode, 15% of it becomes NiO at the sintering
temperature of 1000°C (1300 K). On the contrary, when Cu is used as an electrode, only 0.2% of Cu
becomes Cu2O. This calculation clearly indicates that Cu should be used for the internal electrode
of lead zirconate titanate (PZT)–based capacitor materials. The reader is requested to try similar
thermodynamic calculations on the stability of Ni and Cu for BaTiO3. The results should indicate
the suitability of Ni for BT.
4.3 HYBRID SUBSTRATES

Recent technology has introduced a set of capacitors and resistors as well as conducting leads
onto a hybrid multilayer substrate, without making discrete capacitors and resistors. Figure 4.7
shows a cross-sectional view of a monolithic multicomponent ceramic (MMC) substrate for a

Ni TiO2 Cu TiO2
Pb NiO PbTiO3 Cu2O
NiTiO3 1
Condensed phases (mol ratio)

1
Condensed phases (mol ratio)
0.1
0.1
0.01
0.01 0.001
50 100 150 50 100 150
(a) Temperature (K) (b) Temperature (K)
FIGURE 4.6 Thermodynamic calculation on (a) PbTiO3 + Ni: closed volume and (b) PbTiO3 + Cu: closed volume.
Pod
1 M Tco layer
Conductor
2M
Resistor 3R
4R
5R
Glass-ceramic layer
(low K) 6C
6C
6C
6C
6C
7M
8C
8C
Dielectric ceramic layer
8C
(high K)
8C
N layers
9 M bottom layer
FIGURE 4.7 Cross-sectional view of a monolithic multicomponent ceramic substrate for a voltage-
controlled oscillator.
voltage-controlled oscillator,4 where resistors and capacitors are included in the substrate using
a tape casting technique. Using this MMC substrate, the voltage-controlled oscillator has been
reduced by 1/10 in size (volume).
By spreading out the integration technology, thin-film capacitors will be widely used. Figure 4.8
shows an example with BaSrTiO3 with permittivity ε = 500.5 Figure 4.8a shows a cross-sectional
SEM image, and Figure 4.8b is a high-resolution TEM image of the interface between BaSrTiO3
and the Pt electrode.

Cu BaSrTiO3
Pt
BaSrTiO3
Pt
SiO2
500 nm Pt 6 nm
(a) (b)
FIGURE 4.8 Thin-film capacitor with BaSrTiO3 with permittivity ε = 500. (a) Cross-sectional SEM image
and (b) high-resolution TEM image of the interface between BaSrTiO3 and the Pt electrode. (From Suzuki, T.
et al., ISIF 2005 Proceedings, 2006. With permission.)
4.4 RELAXOR FERROELECTRICS

Relaxor ferroelectrics such as Pb(Mg1/3Nb2/3)O3 and Pb(Zn1/3Nb2/3)O3 have been utilized for very
compact chip capacitors. The reasons why these complex perovskites have been investigated inten-
sively for capacitor applications are
1. Their very high permittivity

2. Their temperature-insensitive characteristics (i.e., diffuse phase transition) in comparison
with normal ferroelectric perovskite solid solutions
However, the relaxors exhibit a problem, namely, dielectric relaxation (frequency dependence of
permittivity), a characteristic highlighted by their name.
4.4.1 HIGH PERMITTIVITY

An intuitive crystallographic model (rattling ion model) has been proposed to explain the giant
permittivity of these disordered perovskites.6 Figure 4.9a and b shows the ordered and disordered
structures for an A(BI,1/2BII,1/2)O3 perovskite crystal. Assuming a rigid ion model, a large “rattling”
space is expected for the smaller B ions in the disordered structure (Figure 4.9b) because the large
B ions prop open the lattice framework. Much less rattling space is expected in the ordered arrange-
ment (Figure 4.9a), where neighbor atoms collapse systematically around the small B ions.
(a) (b)
FIGURE 4.9 Crystal structure models of the A(BI,1/2BII,1/2)O3–type perovskite: (a) the ordered structure with
a small rattling space and (b) the disordered structure with a large rattling space [open circle = BI (lower
valence cation) and solid circle = BII (higher valence cation)].

When an electric field is applied to a disordered perovskite, the B ions (usually high valence
ions) with a large rattling space can shift easily without distorting the oxygen framework. A larger
polarization can be expected for unit magnitude of electric field; in other words, larger dielectric
constants and larger Curie–Weiss constants should be typical in this case. On the other hand, in
ordered perovskites with a very small rattling space, the B ions cannot move easily without distort-
ing the octahedron. A smaller permittivity and a Curie–Weiss constant are expected.
4.4.2 DIFFUSE PHASE TRANSITION

The exact reason why the phase transition is diffuse in the relaxor ferroelectrics has not yet been
clarified. We introduce here the microscopic composition fluctuation model which is one of the
most widely accepted models for the relaxor ferroelectrics.7–9 Within a single Känzig region [the
minimum polar region size in which cooperative polarization (ferroelectricity) can occur], typically
on the order of 10–100 nm, the model applied to a Pb(BI,1/3BII,2/3)O3 relaxor assumes a local fluctua-
tion of the BI2+ and BII5+ ions.
Figure 4.10 shows a computer simulation of the composition fluctuation in an A(BI,1/2BII,1/2)
O3–type crystal calculated for various degrees of ionic ordering. The fluctuation of the BI/BII frac-
tion x obeys a Gaussian error distribution.9 H.B. Krause has reported the short-range ionic ordering
Ordered Number of regions
20
10
0.3 0.4 0.5 0.6 0.7

Fraction of
Number of regions
20
10
0.3 0.4 0.5 0.6 0.7

Fraction of
Disordered
Number of regions
20
10
0.3 0.4 0.5 0.6 0.7

Fraction of
FIGURE 4.10 Computer simulation of the composition fluctuation in an A(BI,1/2BII,1/2)O3–type crystal cal-
culated for various degrees of ionic ordering (Känzig region size in the simulation: 4 × 4). (From Uchino, K.
et al., Jpn. J. Appl. Phys. 20, 171, 1981.)

A A
B
B C
C
110
D
D
E
F G
E G
F
H 2.86 Å
H
I
I
FIGURE 4.11 High-resolution electron-microscope image of a Pb(Mg1/3Nb2/3)O3 single crystal sample with
(110) plane. Note ion-ordered islands in the range of 2–5 nm. (From Krause, H.B. et al., Acta Cryst. A, 35,
1015, 1979.)
of Pb(Mg1/3Nb2/3)O3 observed by electron microscopy.10 The high-resolution image in Figure 4.11

reveals somewhat ordered islands in the range of 2–5 nm, each of which may have a slightly different
transition temperature.
In the case of permittivity, for example, by superimposing the Curie–Weiss law for each cluster
with a different Curie temperature, we obtain a rather broad permittivity peak, which provides a
more stable temperature change. Thus, sometimes the Curie range is specified rather than the Curie
point. The permittivity of relaxor ferroelectrics in the paraelectric region obeys the following qua-
dratic relation6:
1 ε = 1 ε 0 + (T − TC ) C*
2
(4.2)
rather than the normal Curie–Weiss law:
1 ε = (T − TC ) C (4.3)
In order to improve the temperature coefficient of permittivity by promoting a more diffused phase
transition, the following techniques are applied:9
(a) Ion-disordered crystals produced by

(a-1) Adding a nonferroelectric component [e.g., (Pb,Ba)(Zr,Ti)O3, where BaZrO3 is a nonpo-
lar material]
(a-2) The generation of lattice vacancies [e.g., (Pb,La)(Zr,Ti)O3]
(b) Short-range ordering within the crystal due to the generation of cation-ordered clusters
[e.g., Pb(Mg1/3Nb2/3,Ti)O3 and Pb(Mg1/2W1/2,Ti)O3 (PMW–PT)]
Improvement of the temperature coefficient of permittivity by means of (b) is exemplified by the

solid solution PMN–PT incorporating PMW or Ba(Zn1/3 Nb2/3)O3 (BZN). The addition of PMW
tends to generate microclusters of the 1:1 ordered type, and BZN clusters of the 1:2 ordered type.

4.4.3 DIELECTRIC RELAXATION

Another significant characteristic of these “relaxor” ferroelectrics is dielectric relaxation (frequency
dependence of the permittivity), from which their name is derived. The temperature dependence of
permittivity for Pb(Mg1/3Nb2/3)O3 is plotted in Figure 4.12 at various measuring frequencies.11
With increasing frequency, the permittivity in the low-temperature (ferroelectric) phase decreases
and the peak temperature near 0°C shifts towards higher temperature; this is contrasted with the
behavior of normal ferroelectrics such as BaTiO3, where the peak temperature changes little with the
frequency. This is understood to be caused by shallow multipotential wells associated with the locally
distorted perovskite cell due to the disordered ionic arrangement (Skanavi-type dielectric relaxation),12
in addition to a ferroelectric phase transition phenomenon. Figure 4.13 illustrates the model. The
Skanavi-type provides local dipoles and exhibits an “electret”-like property. Of course, when no elec-
tric field is applied initially, no charge or polarization is expected because of noncooperative coupling
in the crystal. When a long-range cooperative phenomenon, that is, dipole coupling (ferroelectricity), is
superposed (pictured by the springs connecting the constituent ions), the net polarization appears.
12000
1
10000 2
3
4
Permittivity ε
8000
5
6000 6
tan δ
6
0.20 4000
0.15 2000
2
0.10
1
5 4
0.05
0
–200 –150 –100 –50 0 50 100 150
Temperature (°C)
FIGURE 4.12 The temperature dependence of the permittivity and tan δ in Pb(Mg1/3Nb2/3)O3 for various
measuring frequencies (kHz): (1) 0.4, (2) 1, (3) 45, (4) 450, (5) 1500, and (6) 4500.
(a) Skanavi-type
(b) Ferroelectric
FIGURE 4.13 Multipotential well model for (a) the Skanavi-type and (b) the ferroelectric relaxors. Note the
difference in the cooperative phenomenon, represented by springs.

α–
α+
– +
ΔU μ
μ
F
FIGURE 4.14 Ion in a double-minimum potential.
EXAMPLE PROBLEM 4.2

Consider an order–disorder-type ferroelectric with ions trapped in the same double-minimum
potential with a relatively low barrier between the two minima (Figure 4.14). Thus, each unit
lattice cell has a dipole moment, and the dipole–dipole coupling is taken into account at a tem-
perature lower than the Curie temperature. Under a quasi-dc field, the ion follows the electric
field alternating between the positive and negative potentials. However, with increasing drive
frequency the ionic motion exhibits a delay with respect to the electric field due to the potential
barrier ΔU. This is an intuitive explanation for the dielectric relaxation in a ferroelectric phase.
1. Using a mathematical representation, derive the Debye dispersion relation for a monodis-
persive case:
ε (ω ) = ε S (1 + jωτ ). (P4.2.1)
2. Also discuss how the above dispersion obeys the so-called Cole–Cole relation (i.e., the real
and imaginary parts of permittivity trace a half circle on a complex permittivity plane).
Solution
When an external electric field E is applied, the local field F in the crystal is described by
F = E + γP. (P4.2.2)
The transition probability for an ion from the − to the + in Figure 4.14, α +, and the opposite tran-
sition probability α−, are expressed as
α + = Γ exp ⎡⎣ − (ΔU − μF ) kT ⎤⎦ , (P4.2.3)
and
α − = Γ exp ⎡⎣ − (ΔU + μF ) kT ⎤⎦ , (P4.2.4)
respectively. Here, ΔU is the barrier height between the two potential minima, μ is the dipole
moment, and Γ is a constant for normalizing the probability.
If we introduce the number of + (or −) direction dipoles per unit volume N+ (or N−), the total
dipole number is given by N = N+ + N−, and the polarization (per unit volume) is represented as
P = (N + − N − )μ. (P4.2.5)
The time dependence will be expressed as
dN + dt = N − α + − N +α − (P4.2.6)

dN − dt = N + α − − N − α + . (P4.2.7)
Then
dP /dt = μ(dN + /dt − dN − /dt )
= 2μ( N − α + − N +α − ), (P4.2.8)
where
N + = (1/2)(N + P /μ), (P4.2.9)
N − = (1/2)(N − P /μ). (P4.2.10)
Suppose that the external electric field E = E 0ejωt is small (PS >> ε0εE) and that the polarization
is given by
P = PS + ε 0 εE0 e jωt (P4.2.11)
From Equation P4.2.8,
Left-side = (jω)ε 0 εE
= Right-side = 2μ{(N − Γexp[ − (ΔU − μF )/kT ] − N+Γexp[ − (ΔU + μF )/kT ]}
= 2μ{(N − Γexp[ − (ΔU − μ(E + γP ))/kT ] − N +Γexp [− (ΔU + μ(E + γP ))/kT ]}
= 2μΓexp( − ΔU /kT ){N − exp[μ(E + γ (PS + ε 0 εE ))/kT ]
−N + exp [− μ(E + γ (PS + ε 0 εE ))/kT ]},
Further using Equations P4.2.9 and P4.2.10,
= 2μΓexp( − ΔU /kT ){(1/2)(N − P /μ)exp(μγPS / kT )exp[μ(1 + γε 0 ε)E /kT ]
− (1/2)(N + P /μ)exp(− μγPS )/kT )exp[ − μ(1 + γε 0 ε)E /kT ].
Taking into account PS >> ε0εE and exp(x) ≈ 1 + x (for X >> 1), the above calculation is trans-
formed into
(jω)ε 0 εE
= 2μΓexp (− ΔU /kT ){(1/2)(N − P /μ)exp(μγPS /kT )[1 + μ(1 + γ ε 0 ε)E /kT ]
− (1/2)(N + P /μ)exp( − μγPS )/kT )[1 − μ(1 + γε 0 ε)E /kT ]
= 2Γexp( − ΔU /kT )[μN{sinh(μγPS / kT ) + μ (1 + γε 0 ε)(E /kT )cosh(μγPS /kT )}
− P{cosh(μγPS /kT ) + μ(1 + γε 0 ε)(E /kT )sinh(μγPS /kT )}] (P4.2.12)
Using the relation P = PS + ε0εE on Equation P4.2.12, we obtain
(jω)ε 0 εE
= 2Γexp( − ΔU / kT )[μN{sinh(μγPS /kT ) + μ(1 + γε 0 ε)(E /kT )cosh(μγPS /kT )}
− PS{cosh(μγPS /kT ) + μ(1 + γε 0 ε)(E /kT )sinh(μγPS /kT )}
− ε 0 εE{cosh(μγPS /kT ) + μ(1 + γε 0 ε)(E / kT )sinh(μγPS /kT )}]

= 2Γexp( − ΔU /kT )[μNμ(1 + γε 0 ε)(E /kT )cosh(μγPS /kT )
− PSμ(1 + γε 0 ε)(E /kT )sinh(μγPS /kT ) − ε 0 εE cosh(μγ PS /kT )]. (P4.2.13)
For the last transformation, we used the relationship of μNsinh(μγPS/kT) − PScosh(μγPS/kT) = 0

and the neglect of the higher order E2. Introducing the static permittivity εS at ω = 0, we obtain
μNμ(1 + γε 0 εS )(E /kT ) cosh(μγPS /kT ) − PSμ (1 + γε 0 ε S )(E /kT ) sinh(μγPS /kT )
− ε 0 ε S E cosh(μγPS /kT ) = 0. (P4.2.14)
Now, for a high-frequency permittivity, from Equation P4.2.13
ε (ω ) = ε S (1 + jωτ ), (P4.2.15)
where
τ = (1 + γε 0 ε S )τ 0 cosh(μγPS /kT ), (P4.2.16)
τ 0 = 1/2Γexp (− ΔU /kT ). (P4.2.17)
The subscript s of εS stands for a static value (ω = 0), and in the paraelectric phase, it should fol-
low the Curie–Weiss law:
ε s = C (T − TC ). (P4.2.18)
Equation P4.2.15 can be rewritten as
ε(ω) = ε ′(ω) + jε″(ω)
where
ε ′ (ω ) = ε s ⎢⎡1+ (ωτ ) ⎥⎤ ,
2
⎣ ⎦
ε ′′ (ω ) = ωτε s ⎢⎡1+ (ωτ ) ⎥⎤ .

2
(P4.2.19)
⎣ ⎦
Τhe so-called Cole–Cole relation is obtained from Equation P4.2.19, which is a half circle with
the following formula (see Figure 4.15):
(ε ′(ω) − ε S 2 )2 + ε ′′(ω)2 = ( ε S 2 )2 . (P4.2.20)
ε˝
ω=∞ ω=0
0 εS/2 εS ε˝
FIGURE 4.15 Cole–Cole plot for a double-minimum potential model.

6 × 104 Depoled
0.3
5 × 104
f Increasing 0.24
Dielectric constant
4 × 104
0.18 15 μm
tan δ
4
3 × 10
0.12
2 × 104 f Increasing
0.06
1 × 104
0 0
20 40 60 80 100 120 140 160 180 200
(a) Temperature (°C)
4 0.3
6 × 10 Poled
5 × 104
0.24
f Increasing
Dielectric constant
4 × 104
15 μm
0.18 tan δ
3 × 104
0.12
2 × 104
f Increasing 0.06
1 × 104
0 0
20 40 60 80 100 120 140 160 180 200
(b) Temperature (°C)
FIGURE 4.16 Dielectric constant and loss vs. temperature measured for (a) an unpoled and (b) a poled
Pb(Zn1/3Nb2/3)O3 single crystal sample [rhombohedral (111) plates]. The domain configurations are also pic-
tured on the right-hand side. When the macrodomains were induced by an external electric field, the dielectric
dispersion disappeared and the loss became very small. Microdomains seem to be the origin for the dielectric
relaxation and the dielectric loss. (From Mulvihill, M.L. et al., Proc. 8th European Mtg. Ferroelectricity,
Nijmegen, 1995.)
It is notable that there is an alternative explanation for the dielectric relaxation, which has been
proposed by Mulvihill et al. for Pb(Zn1/3Nb2/3)O3 single crystals.13 Figure 4.16a and b shows the
dielectric constant and loss vs. temperature for an unpoled and a poled PZN single crystal sam-
ple, respectively. The domain configurations are also pictured. The macroscopic domains were not
observed in an unpoled sample even at room temperature (Figure 4.16a right), only in that state large
dielectric relaxation and loss were observed below the Curie temperature range. Once the macrodo-
mains were induced by an external electric field (Figure 4.16b right), the dielectric dispersion disap-
peared and the loss became very small (i.e., the dielectric behavior became rather normal) below
100°C. As the temperature is increased, the macroscopic domains disappear in the poled sample
at 100°C, then immediately above this temperature both large dielectric dispersion and loss appear

(Figure 4.16b left). Therefore, the dielectric relaxation appears to be associated with the microdo-
mains generated in this material. This model suggests the relaxation origin from the microdomain
clusters, rather than each lattice dipole fluctuation. Mathematical treatment of the relaxor behavior
originating from the presence of microdomains has not yet been conducted.
Chapter Essentials
1. Basic specifications required for capacitors:

(a) Small size, large capacitance
(b) High-frequency characteristics
(c) Temperature characteristics
2. Electrostatic capacitance of an MLC with the same volume (layer number = n):
C = n2 (ε0εS /L ).
3. Characteristics of relaxor ferroelectrics:

(a) High permittivity
(b) Temperature-insensitive characteristics (i.e., diffuse phase transition)
(c) Dielectric relaxation
4. Modified Curie–Weiss law for relaxor ferroelectrics:
1 ε = 1 ε0 + (T − TC )2 C*.
5. Dielectric relaxation is explained conventionally by the Debye dispersion model based on
the double potential minima on each crystal lattice.
6. Dielectric relaxation in some relaxor ferroelectrics is attributed to the presence of microdo-
mains. Once macrodomains are induced by an external electric field, the dielectric disper-
sion disappears and the loss becomes very small.
Chapter Problems
4.1 An MLC (100 layers) is made from a 3 μm thick sheet with a dielectric material ε =
3000. Assuming a 90% ratio of overlapped electrode area over the chip surface area,
calculate the chip area to obtain a total capacitance of 10 μF.
4.2 When the relaxation time is distributed, the permittivity dispersion follows as
ε (ω) = εS ⎡⎢1 + ( jωτ ) ⎤⎥

β
⎣ ⎦
where β < 1. Discuss the Cole–Cole plot change in comparison with the case β = 1 in
Figure 4.15 (monodispersion model).
Check Point
1. (T/F) X7R-labeled capacitors are composed of the high dielectric constant K (>10,000)
with the capacitance deviation less than ±15% in a standard operation temperature range
(from −30°C up to 120°C). True of False?
2. (T/F) There are 1-layer and 10-layer capacitors with the same total thickness and the area.
The 10-layer capacitor has 10 times higher capacitance in comparison with the single-layer
capacitor. True or False?
3. (T/F) Cu internal electrode is suitable for BT-based multilayer chips (MLCs), while Ni is
suitable for PMN-based MLCs. True or False?

(a) (b)
FIGURE 4.17 Crystal structure models of the A(BI,1/2BII,1/2) O3-type perovskite: (a) the ordered structure
with a small rattling space and (b) the disordered structure with a large rattling space [=BI (lower valence
cation) and =BII (higher valence cation)].
4. We consider two crystal structure models of the A(BI,1/2BII,1/2)O3-type perovskite: (a)

ordered structure and (b) the disordered structure, as shown in Figure 4.17. Which struc-
ture is expected to exhibit a larger permittivity?
5. (T/F) The relaxor ferroelectric shows a decrease in permittivity on increasing the measur-
ing frequency. True or False?
6. (T/F) When we plot the real and imaginary parts of permittivity on a 2-D map (Cole–Cole
plot), the plots follows on an exact half circle for all relaxor ferroelectrics. True or False?
REFERENCES
1. Murata Catalog: Miracle Stones.
2. K. Utsumi: Private communication at 4th US–Japan Seminar on Dielectrics & Piezoelectric Ceramics
(1989).
3. H. Kishi: Mater. Integr., 19(3), 47–51 (2006).
4. K. Utsumi, Y. Shimada, T. Ikeda, and H. Takamizawa: Ferroelectrics, 68, 157 (1986).
5. T. Suzuki, K. Morito, and Y. Iwazaki: ISIF 2005 Proceedings (2006).
6. K. Uchino, L. E. Cross, R. E. Newnham, and S. Nomura: J. Phase Transit., 1, 333 (1980).
7. W. Känzig: Helv. Phys. Acta, 24, 175 (1951).
8. B. N. Rolov: Fiz.Tverdogo Tela, 6, 2128 (1963).
9. K. Uchino, J. Kuwata, S. Nomura, L. E. Cross, and R. E. Newnham: Jpn. J. Appl. Phys., 20(Suppl. 20–4),
171 (1981).
10. H. B. Krause, J. M. Cowley, and J. Wheatley: Acta Cryst. A, 35, 1015 (1979).
11. G. A. Smolensky, V. A. Isupov, A. I. Agranovskaya, and S. N. Popov: Sov. Phys. Solid State, 2, 2584
(1961).
12. G. I. Skanavi, I. M. Ksendzov, V. A. Trigubenko, and V. G. Prokhvatilov: Sov. Phys. JETP 6, 250
(1958).
13. M. L. Mulvihill, L. E. Cross, and K. Uchino: Proc. 8th European Mtg. Ferroelectricity, Nijmegen
(1995).

5 Ferroelectric Memory Devices
Extremely large-scale integrated semiconductor memories using ferroelectric films have been inves-
tigated recently. Since conventional Si micromachining technology coupled with silicon oxide/nitride
and metal, is limited in its ability to produce finer scale capacitors, utilization of ferroelectrics having
high permittivity or polarization hysteresis has been considered as a possible solution to make large-
scale memories.
5.1 DRAM
5.1.1 PRINCIPLE OF DRAM
There are volatile and nonvolatile memory devices in erasable semiconductor memories. Dynamic
random access memory (DRAM), which is widely used because of its high integration capability,
is an example of volatile memory. Data stored in these memories are lost when the electric power
is shut off. On the contrary, nonvolatile memories include a circuit-latch multiple field effect tran-
sistors (FETs) and Si-surface-potential control metal-oxide-semiconductor (MOS) FET channel.
However, both types, in general, have problems in integration density and writing time.
Figure 5.1 shows the fundamental structure of a DRAM composed of a transistor and a capacitor;
a SiO2 film capacitor is connected to the source of a MOSFET. Figure 5.2 shows a realistic structure
of the DRAM. During the writing process, one DRAM element is chosen by x–y addressing; that
is, voltage is applied on both the gate and the drain electrodes simultaneously, thus accumulating
charge on the SiO2 film capacitor (memorizing). Since the accumulated charge leaks, the capacitor
must be recharged repeatedly (refreshing).
The electron–hole pair generation around the FET by the IC package or natural radiation changes
the amount of charge on the capacitor, and sometimes destroys the memory (soft error). In order to
retain memory, the capacitance of the memory capacitor must be higher than 30 fF (remember f =
10 −15 and called “femto”).
EXAMPLE PROBLEM 5.1

Explain the generation process of the depletion and inversion layers in a MOS structure (p-type
Si) using a simple energy band model as pictured in Figure 5.3, when a positive voltage is applied
on the metal. Describe the Fermi level in the metal region and the hole and electron concentra-
tions using + and − symbols in the energy band model. For simplicity, you can use the assump-
tion that the flat band voltage is close to zero.
Solution
Figure 5.4 shows the energy band change of a MOS structure (with a p-type semiconductor)
under an applied voltage.
(a) When the gate is forward-biased, holes will accumulate in the semiconductor region near
the oxide.
(b) When the gate is reverse-biased, a depletion layer (i.e., a region depleted of holes) is
generated.
(c) When the applied voltage exceeds a certain threshold voltage (VT), an inversion layer
(an electron-rich region) is generated. This condition is given by: semiconductor surface
voltage ψs > ψB, where ψB is the difference between the intrinsic Fermi level (Ei) and the
Fermi level (EF) [ψB = (kT/q) ln(Na /ni)].
131

Word line
MOSFET Gate
Source Drain
Fixed voltage
Capacitor
Bit line
FIGURE 5.1 Fundamental structure of a DRAM, composed of a MOSFET and a capacitor.
Poly -Si word line Al evaporated

bit line
SiO2
SiO2
n+ n+
Capacitor MOSFET
Poly-Si p-type Si
fixed voltage
FIGURE 5.2 The structure of a DRAM.
Conduction band
Fermi level Intrinsic level

EF EF
Valence band
Metal M
Semiconductor (p-type)
Oxide O
FIGURE 5.3 Energy band model for a MOS structure.
Based on Example Problem 5.1 we learned how to generate the inversion layer, now we will learn
how we can address the memories (only when both the gate and drain is biased, the drain current
flows) by the following example, which is the basic principle to understand Figure 5.1.
EXAMPLE PROBLEM 5.2

Let us consider an n-channel enhancement-mode MOSFET as illustrated in Figure 5.5. A posi-
tive gate voltage induces the electron inversion layer, which then connects the n-type source and
the n-type drain regions. Discuss the drain current behavior as a function of the drain/source
voltage.1
Solution
A positive gate voltage induces the electron inversion layer, which then connects the n-type
source and drain regions. The source terminal is the source of carriers that flow through the
channel to the drain terminal. In such an n-channel device, electrons travel from the source to
the drain so that the conventional current flows from the drain to the source.

Ferroelectric Memory Devices 133
M O S
Electron
E1 0
EF Hole
(a) Forward bias
Metallic electrode
E2
SiO2
E2 0 EF
Depletion layer
(b) Depletion state
p-type semiconductor
E3 > E2
EF
E3 > 0 Inversion layer
(c) Inversion state
FIGURE 5.4 Energy band change of a MOS structure (with a p-type semiconductor) under an applied
voltage.
Gate
SiO2 film
Source n+ Drain n+
p-type semiconductor
FIGURE 5.5 MOSFET fabricated with a p-type semiconductor (n-channel enhancement-mode MOSFET).
Remember that the depletion layer has low conductivity, which is analogous to a shielded cable,
a lead wire (the inversion layer) covered by an insulating coat (the depletion layer). This is analo-
gous to water flowing in a tube, where the water (the electron) flows in a tube (the inversion layer)
surrounded by rubber material (the depletion layer). When the tube is pinched off, the water flow
is restricted.
Making the assumption that the flat band voltage is close to zero, the application of the gate
voltage EG easily creates the inversion layer (Figure 5.6a). When a small drain voltage (EDS <
EG) is applied, the electrons in the inversion layer will flow from the source to the positive drain
terminal. Since for small EDS, the channel region has the characteristics of a conducting lead,
we expect
I D = gd EDS . (P.5.2.1)

EG ED
Gate
Source n+ Drain n+
Inversion layer
Electron flow Depletion layer
(n-channel)
(a) Drain voltage ED < gate voltage EG
Gate
Source n+ Drain n+
Electron flow Inversion layer

Depletion layer
(b) Drain voltage ED = gate voltage EG
Gate
Source n+ Drain n+
Depletion layer
Electron flow Inversion layer
(c) Drain voltage ED > gate voltage EG
FIGURE 5.6 Change in the n-channel with the drain/source voltage for an n-channel enhancement mode
MOSFET.
When EDS is increased to the point where the potential drop across the oxide at the drain terminal
is equal to zero (precisely speaking, equal to the threshold voltage VT), the induced inversion
charge density is zero at the drain terminal. This effect is schematically shown in Figure 5.6b. At
this point (EDS = EG), the incremental conductance at the drain becomes zero. The slope of the
ID vs. EDS curve becomes zero.
When EDS becomes larger than the above value (EG), the point in the channel at which the
inversion charge is just zero shifts toward the source terminal (Figure 5.6c). In this case, elec-
trons enter the channel at the source, travel through the channel toward the drain, and then, at
the pinch-off point the electrons are injected into the space charge region (depletion layer) where
they are swept by the E-field to the drain contact. If we assume that the change in channel length
is small compared to the original length, the drain current becomes constant for EDS > EG. This
region is referred to as the saturation region. Figure 5.7 shows the ID vs. EDS curves.
ED (sat) = EG–VT EG3 > EG2

Drain current ID
EG2 > EG1
EG1 > VT
Drain/source voltage ED
FIGURE 5.7 ID vs. EDS curves for various EG.

5.1.2 FERROELECTRIC DRAM

Since the conventional SiO2 film plane structure cannot maintain sufficient capacitance with decreas-
ing element area, a multilayer type and a trench type (normal hole structure on a Si substrate) structured
capacitors have been proposed. Yet even these complex three-dimensional structures have their own
limitations. Alternatively, ferroelectric materials with their high permittivities are attractive candidates
for DRAM devices.2 The DRAM application in this section uses the paraelectric phase. Candidate
materials for this application are ferroelectrics with Curie temperatures below room temperature.
Because the dielectric constant of ferroelectrics is typically above 1000, compared to SiO2’s
3.9, using a ferroelectric thin film with the same thickness as a SiO2 film, the capacitor size to
obtain 30 fF is significantly decreased. The size gain is by 1/250 in terms of the area, or by 1/16 in
linear scale. Hence, much higher memory density is attainable for a ferroelectric DRAM device.
Of course, the situation is not as simple as stated, because the thickness of the ferroelectric film
required to exhibit a sufficiently large permittivity sometimes needs to be thicker than the SiO2 film
typically used in the conventional MOS structure.
The material first studied for this purpose was strontium titanate (SrTiO3), a ferroelectric that
does not exhibit hysteresis in its polarization vs. electric field curve or at the permittivity peak in
its temperature characteristics, and has a dielectric constant of around 300 at room temperature.
Note that for DRAM application, a high dielectric constant is essential, but the typical ferroelectric
dielectric hysteresis must be avoided; this is accomplished by utilizing the paraelectric phase by
operating just above the Curie temperature.
The SrTiO3 film is superior to the conventional ferroelectrics from a manufacturing point of view.
Although conventional ferroelectric thin films tend to exhibit a decrease in their dielectric constant
with decreasing thickness, below 200 nm, SrTiO3 films do retain a relative permittivity of about 220
down to a thickness of 50 nm, as shown in Figure 5.8.3 Therefore, a 50 nm film of SrTiO3 (ε = 220)
is equivalent to a 0.88 nm film of SiO2 with ε = 3.9.
EXAMPLE PROBLEM 5.3
(a) Calculate the size of the square-shaped capacitor for 30 fF with a 50 nm thick SiO2 film
(ε = 3.9).
(b) Calculate the size of the square-shaped capacitor with the same thickness SrTiO3 film
(ε = 220).
10 kHz
300
540ºC
Permittivity
200
430ºC
100
Pd (500 nm)/sapphire sub.

0
0 50 100 150 200
Film thickness (nm)
FIGURE 5.8 Film thickness dependence of the dielectric constant in SrTiO3.

Solution
(a) Based on the capacitance formula, C = ε0ε(S/t)
30 × 10 −15 [F] = 8.856 × 10 −12 × 3.9 × (S/50 × 10 −9 [m]). (P5.3.1)
Thus
S = 43.4 × 10 −12 [m 2 ], or L = 6.6[μm].
(b) Similarly
30 × 10 −15 [F] = 8.856 × 10 −12 × 220 × (S/50 × 10 −9 [m]). (P5.3.2)
S = 0.770 × 10 −12 [m 2 ], or L = 0.88[μm]
6.6 μm L square for SiO2 can be reduced to 0.88 μm L square, when SrTiO3 is adopted, replac-
ing the conventional SiO2 capacitor. This considerable decrease in size accelerated “extremely
large-scale integration.”
The solid solution Ba xSr1−xTiO3 [BST] between SrTiO3 and BaTiO3 has also been investigated
because of its larger dielectric constants at room temperature, which can further improve the capaci-
tor integration. The SiO2 equivalent thickness of 0.47 nm was reported using BST films.4
The DRAM capacitor films must have high resistivity in order not to leak the accumulated charge
during the refreshing cycle. A leakage current less than 10 −7 A/cm2 has been obtained for SrTiO3 and
Ba xSr1−xTiO3 films, which satisfies the capacitance requirement for 256 Mb level devices. General
requirements for DRAM capacitor thin films are
1. High dielectric constant as a thin film

2. Low leakage current
3. Micromachinability
4. Low diffusion into the semiconductor substrate
5. Low contamination during the fabrication process
The dry etching technique has been successfully applied for micromachining of ferroelectric films,
and diffusion and contamination problems have been solved by decreasing the film fabrication
temperature and by carrying out the ferroelectric fabrication as late as possible in all the processes.
Thus, compatibility with the semiconductor process has been established, and prototype 256 Mb
level DRAMs have been fabricated and their functions were verified.
5.2 NONVOLATILE FERROELECTRIC MEMORY

5.2.1 FERAM (INVERSION CURRENT TYPE)
When a ferroelectric thin film with a large polarization–electric field hysteresis is used as the
memory capacitor in the structure pictured in Figure 5.1, a nonvolatile memory is realized. When a
voltage is applied to the gate and the FET assumes the “on” state, a pulse voltage to the drain gener-
ates a drain current dependent on the remanent polarization state.
Let us assume a P–E hysteresis loop of the ferroelectric film as illustrated in Figure 5.9, and that
the polarization state is on A. When a step voltage is applied and the state shifts to B, the current
flows according to the polarization difference between A and B. On the contrary, when the polar-
ization state is on C first, the current increases dramatically because of the associated spontaneous
polarization reversal. Figure 5.10 shows the current responses to a series of pulses (two positive

Polarization P
B
A
Electric field
FIGURE 5.9 Polarization vs. electric field curve for a ferroelectric film.
4000
PZT (Ti = 60%)
Film = 282 nm
Current density (A/cm )
2400 Ipost
2
20 × 20 μm2
Iup
800
–800
Idown Ppost Pup
–2400
Pnega Pdown
Inega
–4000
0 15 30 45 60 75
Time (ns)
FIGURE 5.10 Current responses to a series of pulses (two positive pulses followed by two negative pulses)
on a PZT film with 20 × 20 μm2 electrodes.
pulses followed by two negative pulses) on a lead zirconate titanate (PZT) film with 20 × 20 μm2
electrodes.5 When a positive pulse is applied just after the negative pulses, a large current Iposi is
observed, which includes the polarization reversal. However, the second positive pulse generates
only a small current Iup. Thus, the observed current amount for a positive pulse can indicate the ini-
tial polarization state; that is, an on or off state, or 1 or 0 state. In this memory device, after reading
the initial state by applying the positive voltage, the minimum polarization state becomes A for all
the times; that is, the reading process is destructive. Hence, in order to retain the memory state, a
writing process similar to the case of DRAM is required every time.
Because a large electric field is applied on a ferroelectric film at every reading process in the fer-
roelectric RAM (FeRAM), as discussed above, the polarization hysteresis characteristic degrades
with increasing cycles. This problem of ferroelectric films needs to be overcome for nonvolatile
memory applications. From a practical point of view, a lifetime (i.e., the time until the polarization
degradation is observed) of more than 1015 cycles is required.
There are three categories of the origin of characteristic degradation: “fatigue,” degradation in
PS, and “in-print,” which are schematically illustrated in Figure 5.11. Fatigue is characterized by the

Polarization P
Polarization P
Polarization P
1 1 1
Electric Electric Electric

field E field E field E
0 0 0
(a) (b) (c)
FIGURE 5.11 Three categories of the origin of characteristic degradation: (a) “fatigue,” (b) degradation in
PS, and (c) “in print.”
degradation in the polarization magnitude in a dynamic cycle. “Degradation in PS” is the PS reduc-
tion in a static situation. To the contrary, “in-print” is characterized by the polarization curve shift
along the horizontal axis due the internal bias field creation.
The possible origins for the characteristic degradation are related to the ferroelectric film mate-
rial quality (such as crystallinity, generation of oxygen vacancies, and the diffusion of ions) and
interface/junction between the piezofilm and electrode. Much effort has been made to remedy these
problems. The proposed ideas can be categorized as follows:
1. Improvement of the film fabrication process

2. Search for new ferroelectric materials
3. Improvement of electrode materials
Recent thin film materials include layer-structure ferroelectrics. The so-called Y1 material
patented by Symetrix, which has a basic composition of BiSr2Ta2O9, shows superior antifatigue
properties. Figure 5.12 shows the fatigue characteristics for rewriting the remanent polarization
in Y1 and PZT films.6 The magnitude of the remanent polarization does not change significantly
in Y1 even after testing for 1012 cycles, an improvement as compared to the lifetime of 107 cycles
for PZT.
New electrode materials RuO2 and Ir have been found to exhibit improvement in fatigue in com-
parison with the conventional Pt electrode. Furthermore, new drive modes such as a combination of
the DRAM operation during the switch-on stage and the memory mode during the switch-off stage
have been proposed.
Remanent polarization
1.0
Y1
0.8
0.6
PZT
0.4
0.2
0.0
102 104 106 108 1010 1012 1014
Number of pulse repetition
FIGURE 5.12 Fatigue characteristics for rewriting the remanent polarization in Y1 and PZT films.

5.2.2 MFSFET
Figure 5.13 shows a structure of an metal-ferroelectric-semiconductor (MFS) FET, in which a
ferroelectric film replaces a conventional gate oxide (SiO2 insulator) film. The surface potential of
the channel changes, according to the polarization hysteresis of the ferroelectric film, leading to a
change in the carrier amount and in the current.
Historically, the prototype was designed on a bulk ferroelectric, on which a semiconductor film
was deposited and a FET fabricated. Then, the structure was modified to one similar to the present
type, that is, the ferroelectric film is deposited after making a FET on a silicon crystal. Figure 5.14
shows the drain current vs. gate voltage relation in an MFSFET with a PbTiO3 film fabricated on a
SiO2/Si substrate.7 Due to the polarization hysteresis, the drain current exhibits two states, on and
off. Note that because a p-channel is generated on an n-type semiconductor, a negative gate voltage
provides a drain current (on state).
Though the original devices had problems in fatigue and bistability characteristics, this MFSFET
structure is one of the ideal types, because no large electric field is applied on the ferroelectric film
during the reading process, nor is a rewriting process required after reading, in contrast to the con-
ventional FeRAM. In addition, the polarization density required to control the Si surface potential
is relatively small. Thus, in general, the requirements for the ferroelectric film are greatly reduced
in this design.
However, further investigations are required to fabricate a high-quality, nicely oriented ferro-
electric film on a semiconducting Si or a SiO2 film. Recent studies include PbTiO3 and PZT film
applications prepared by laser ablation or MOCVD, a CaF2, SrF2, or CeO2 film on a Si substrate,
and a metal-ferroelectric-MOS (MFMOS) structure with a ferroelectric film deposited on an Ir film
on SiO2/Si.
Gate SiO2
Ferroelectric film
Source p+ Drain p+
n-type silicon
FIGURE 5.13 Structure of an MFSFET.
FIGURE 5.14 Drain current vs. gate voltage relation in an MFSFET with a PbTiO3 film fabricated on a SiO2/
Si substrate. (From Matsui, Y. et al., Proc. 2nd Mtg. Ferroelectric Mater. Appl., Kyoto, Japan, 239, 1979.)

The development of the ferroelectric memory started with DRAMs, then moved into FeRAMs,
and is now focused on MFSFETs. Development of FeRAM began in the late 1980s. Work was
done in 1991 at NASA’s Jet Propulsion Laboratory on improving methods of readout, including a
novel method of nondestructive readout using pulses of UV radiation. FeRAM volume production
at Fujitsu began in 1999. FeRAMs at 1 Mbit density were available in high volume in 2006 from
both Fujitsu and Ramtron.8 Texas Instruments has collaborated with Ramtron to develop FeRAM
test chips in a modified 130 nm process. Fujitsu and Seiko-Epson were in 2005 collaborating in the
development of a 180 nm FeRAM process.
“Flash memory” sticks are widely used nowadays. What are the merits/demerits of FeRAM, in
comparison with Flash?
Answer
FeRAM is one of the alternative nonvolatile memory technologies that offer the same functionality
as Flash memory.
As schematically shown in Figure 5.15, the principle of Flash is the usage of a floating gate. Zero
charge state (a) in the floating gate corresponds to “1” state, and negative charge state (b) corre-
sponds to “0” state. Applying a high voltage (12–18 V) on the control gate and injecting or extrud-
ing the electrons in the floating gate via the tunneling current. Due to this operation principle, the
speed is slow. On the contrary, Figure 5.16 shows the FeRAM operation principle. Applying a volt-
age (1–5 V) on the ferroelectric film and changing the polarization direction. Since the polarization
reversal can be induced by lower voltage drive, quicker response can be attained.
Advantages of FeRAM over Flash and DRAM include: (1) lower power usage, (2) faster write
speed, and (3) a much greater maximum number (exceeding 1016 for 3.3 V devices) of write–erase
cycles. FeRAM’s disadvantages are (4) much lower storage densities than Flash devices, (5) storage
capacity limitations, and (6) higher cost. Table 5.1 summarizes these comparisons.
Flash memory represented 93% of the total nonvolatile memory market, $28.2 billion in the
world in 2007. However, FeRAM market will gradually increase because of their superior perfor-
mance in comparison with the Flash memory in principle.
CG + + + CG + + +
FG FG – – –
n n n n
On Off
p p
(a) (b)
CG + + +
FG
e e
n n
(c)
FIGURE 5.15 Flash memory: (a) “1 State,” (b) “0 State,” and (c) writing/erasing process.

+ + + – – –
P
P + + +
– – –
(a) (b)
FIGURE 5.16 FeRAM: (a) “1 State” and (b) “0 State.”
TABLE 5.1
Comparison of FeRAM Performances with
Flash and DRAM
FeRAM FLASH DRAM
Memory character Nonvolatile Nonvolatile Volatile
Writing time 100 ns 15 μs 100 ns
(Erase 10 ms)
Reading time 100 ns 120 ms 100 ns
Writing voltage 1–5 V 12 V 1–5 V
Rewriting lifetime 1012–1016 105 1016
Chapter Essentials
1. Ferroelectric memory development trend:

DRAM → FRAM → MFSFET
2. Volatile memory DRAM is composed of a FET and a memory capacitor. Capacitor mate-
rial shifted from SiO2 to higher permittivity ferroelectrics (in paraelectric phase).
3. Minimum memory capacitance is about 30 fF. (f = 10 −15)
4. FeRAM is an inversion current type of reading device.
5. MFSFET is a channel surface potential control type of FET.
6. Three categories of the origin of characteristic degradation in ferroelectric memories:
“fatigue,” degradation in PS, and “in print.”
7. Advantages/disadvantages of FeRAM over Flash: (1) lower power usage, (2) faster
write speed, (3) a greater maximum number (exceeding 1016 for 3.3 V devices) of write–erase
cycles, (4) lower storage densities, (5) storage capacity limitations, and (6) higher cost.
Check Point
1. (T/F) A DRAM is composed of a MOSFET and a capacitor. The states “0” and “1”
correspond to “zero charge” or “charged” capacitor states. True or False?
2. (T/F) An n-channel enhancement mode MOSFET becomes “on” only when positive
voltage is applied to both word and bit lines. True or False?
3. Provide three advantages of FeRAM over Flash memory.
4. (T/F) Pt is one of the best electrode materials for a ferroelectric PZT film in FeRAM
devices. True or False?
5. (T/F) Flash memory requires larger energy in writing, in comparison with a FeRAM. True
or False?

80
60 PZT(100)
40 PZT(111)
Polarization μC/cm2
20
–20
–40
–60
–80
–400 –200 0 200 400
Electric field kV/cm
FIGURE 5.17 Polarization vs. electric field curves for [100]- and [111]-oriented films of PZT 70/30. Notice
smaller EC and sharper square curve shape for the [100]-oriented PZT film.
Chapter Problems
5.1 Surveying the recent publications and patents, discuss and summarize the studies on
ferroelectric thin films from the following viewpoints.
(1) List the papers (minimum 5) which report on epitaxially grown PZT films.
(2) Tabulate the experimentally obtained physical parameters of the PZT films, and
compare with the data for bulk ceramics.
(3) Discuss the above deviation briefly with reference to the papers’ results and
conclusions.
(4) Discuss the crystal orientation dependence of the physical parameters of the PZT
films by referring to the paper, Du, X. H., U. Belegundu, and K. Uchino, “Crystal
orientation dependence of piezoelectric properties in PZT: Theoretical expectation
for thin films,” Jpn. J. Appl. Phys., 36[9A], 5580–5587, 1997. Referring to the P vs.
E curves shown in Figure 5.17 for [100]- and [111]-oriented films of PZT 70/30,
discuss which film should be more suitable to the FeRAM applications.
5.2 We learned in Chapter 4 that lead magnesium niobate (PMN)-based ceramics exhibit
very high dielectric constants. If we can fabricate a high-quality thin film of PMN, is it
applicable to the DRAM for microcomputer applications? Discuss the feasibility of this
proposal, taking into account the operation frequency of the microcomputer.
REFERENCES
1. D. A. Neamen: Semiconductor Physics and Devices, 2nd edn., Irwin, Boston, MA (1997).
2. M. Okuyama: Ferroelectric memory, Bull. Ceram. Soc. Jpn., 30(6), 504 (1995).
3. S. Yamamichi, T. Sakuma, K. Takemura, and Y. Miyasaka: Jpn. J. Appl. Phys., 30, 2193 (1991).
4. T. Sakaemori, Y. Ohno, H. Ito, T. Nishimura, T. Horikawa, T. Shibano, K. Sato, and T. Namba: Nikkei
Micro Devices, (2), 99 (1994).

5. T. Mihara, H. Watanabe, C. A. Pas de Araujo, J. Cuchiaro, M. Scott, and L. D. McMillan: Proc. 4th Int.
Symp. Integrated Ferroelectrics, Monterey, CA, p. 137, March (1992).
6. H. Fujii, T. Ohtsuki, Y. Uemoto, and K. Shimada: Jpn. Appl. Phys. Mtg. Appl. Phys. Electron., 456, AP
942235, 32 (1994).
7. Y. Matsui, H. Nakano, M. Okuyama, T. Nakagawa, and Y. Yamakawa: Proc. 2nd Mtg. Ferroelectric
Mater. Appl., Kyoto, Japan, p. 239 (1979).
8. http://en.wikipedia.org/wiki/Ferroelectric_RAM

6 Pyroelectric Devices
The pyroelectric effect in certain materials was recognized a long time ago, and such materials were
referred as “electric stones.” It was observed when such a stone was thrown in the fire, it started to
generate electric charges and caused “cracking” sounds. This is basically due to the temperature
dependence of the spontaneous polarization of a polar material. The pyroelectric sensors are widely
used for monitoring temperature or infrared (IR) radiation. The converse effect is called electroca-
loric effect, which may be a future cooling system.
6.1 PYROELECTRIC MATERIALS

6.1.1 PYROELECTRIC EFFECT
Practical applications of the pyroelectric effect in temperature sensors and IR light detectors have
been enabling some commercial marketing of ferroelectric ceramics.
Figure 6.1 shows comparison of the sensitivity D* of various commercially available IR detec-
tors when operated at the indicated temperature.1 Chopping frequency is 1000 Hz for all semi-
conductor detectors, except the thermopile (10 Hz), thermocouple (10 Hz), thermistor bolometer
(10 Hz), Golay cell (10 Hz), and pyroelectric detector (10 Hz). Each detector is assumed to view a
hemispherical surround at a temperature of 300 K. Theoretical curves for the background-limited
D* for ideal photovoltaic and photoconductive detectors and thermal detectors are also shown.
Triglycine sulfate (TGS), a ferroelectric, represents the pyroelectric sensitivity in this figure.
As shown, the merits of pyrosensors as compared to semiconducting IR-sensor materials are sum-
marized as follows:
1. Wide range of response frequency.

2. Use at room temperature.
3. Quick response in comparison with other temperature sensors.
4. High-quality (optical-grade homogeneity, etc.) materials are NOT necessary for
pyrosensors.
The principle on which the pyroelectric effect is based concerns the charge generation associated
with the spontaneous polarization change with temperature:
j = −(∂PS /∂t ) = −(∂PS /∂T )(∂T /∂t ) = p(∂T/∂t ). (6.1)
Here p (=|∂PS/∂T|) is denoted as the pyroelectric coefficient. The negative sign (-) comes from the
fact (∂PS/∂T) < 0 to define p as positive. The phenomenon is illustrated schematically in Figure
6.2. Two typical electrode arrangements for pyrosensors are illustrated in Figure 6.3a and b. The
former type has higher efficiency because of a wide charge collecting area, but requires a sophis-
ticated fabrication process for applying uniform transparent electrodes for the IR light. Note that
common transparent materials are transparent to the human visible light, but not necessarily to
IR light.
145

1012
77 K Ideal photovoltaic 2π FOV
(PC)
InAs 300 K background
1 93 K InAs (PV) 193 K
(PC)
PbS HgCdTe (PC) 77 K Ideal photoconductor
C) 77 K Ideal thermal detector
1011 PbS (P
5 K HgCdTe (PV) 77 K
) 29
(PC 193 K )77 K
PbS K
(PC) e (PV
D* (cm Hz½ W–1)
) 77 PbSnT
(PV PbSe
b ) 28 K
InS K g (PC K
) 77 77 K Ge:H ) 4.2
1010 (PC PC) 4.2 K (PC
b e ( ( PC) Zn
InS PbS e:Cu G e :
G
Si:As (PC) 4.2 K
Golay cell
5K
V ) 29
s (P Radiation thermocouple
109 InA ) 295
K TGS Pyro
e (PC GaAs QWIP
PbS 77 K
InSb (PEM) 295 K Ge:Au (PC)77 K
Thermistor bolometer
Thermopile
108
1 1.5 2 3 4 5 6 7 8 9 10 15 20 30 40
Wavelength (μm)
FIGURE 6.1 Comparison of D * of various commercially available IR detectors when operated at the indi-
cated temperature. Each detector is assumed to view a hemispherical surround at a temperature of 300 K.
(From Rogalski, A., Sens. Mater., 12, 233, 2000. With permission.)
Heat sensor
Infrared irradiation
–
–
– – – –
+ + + +
PS
+
– – – – +
+
+ + +
FIGURE 6.2 Principle of a pyroelectric sensor: a temperature increase due to the IR irradiation (such as
human body) → spontaneous polarization decrease → variation in electric charge (or current).
PS
PS
Radiation
Radiation
Electrode
area Electrode
area
a
(a) a (b)
FIGURE 6.3 Typical geometric configurations for pyroelectric detectors: (a) face electrodes with the polar-
ization direction parallel to the IR irradiation, and (b) edge electrodes with the polarization direction perpen-
dicular to the irradiation.

Pyroelectric Devices 147
Light intensity
Time
Temperature rise
Time
Pyroelectric current
Time
FIGURE 6.4 Pyroelectric response to chopped IR irradiation.
EXAMPLE PROBLEM 6.1

When a chopped IR beam is incident on a pyroelectric material, what is the wave form of the
induced pyroelectric current?
Solution
When an IR beam is incident on a pyroelectric material, the temperature of the sample will be
increased exponentially according to (1 − e−t/τ). Since the pyroelectric current j is proportional to
(∂T/∂t) (see Equation 6.1)
j ∝ e −t / τ . (P6.1.1)
It is evident from this relation that one pulse of IR light will provide only one pulse of current.
Because the current becomes zero after a certain time, we cannot distinguish whether the IR
beam is coming or not. Therefore, to measure the temperature of an IR irradiating object (such as
meat in a microwave oven), the IR beam from the object needs to be chopped periodically.
When the illumination is periodic, a periodic variation in the temperature of a pyroelectric
sensor will induce a periodic variation in the pyroelectric current, as illustrated in Figure
6.4. Since the current is alternating, rectification is necessary to obtain the light intensity (or
the object temperature). Higher the object temperature, higher the pyrocurrent amplitude.
However, special attention is required for determining the chopping frequency, which is
directly related with the temperature rise time constant τ (a function of the device configura-
tion, as well as the specific heat). When the chopping period is much longer than τ, the current
increases with chopping frequency ω, since the current maximum is already maintained. On
the contrary, when the chopping period is much shorter than τ, the current will not reach to the
ideal maximum value, and the current will decrease on increasing the frequency ω. Therefore,
we can expect the maximum current or sensitivity at a certain chopping frequency, which
is related with the time constant τ. Chopping frequency for regular pyrosensors is around
1~100 Hz.
Lab Demonstration 6.1: Pyrocurrent
1. Prepare a pyroelectric component such as a lead zirconate titanate (PZT) ceramic disk, a
multilayer, etc. If you do not have them, you may find a bimorph speaker in your old elec-
tronic equipment (a mobile phone or a laptop computer). Then, prepare a voltage/current
meter and a hair dryer (heater) (Figure 6.5).
2. Connect a positive (red color) and a negative (white/black color) lead wire of the piezo-
component to a positive/negative connector of the voltage meter. The lead wire color
corresponds to the original poling direction of the PZT ceramic.

Multilayer
piezocomponent
FIGURE 6.5 Pyroelectric effect demonstration with a pyroelectric component, a hair dryer, and a voltage/
current meter.
3. Using the hair dryer, increase/decrease the temperature of the piezocomponent, and dem-
onstrate positive/negative voltage/current, respectively.
4. Using an unidentified (no information on the polarization direction) piezocomponent, ask
the students how to identify the polarization direction.
6.1.2 RESPONSIVITY
When the incident light power flux is W exp(jωt) (i.e., chopped IR irradiation with a frequency ω),
the amplitude of the temperature variation is provided by2
( )
−1/2
ΔT = ηWA γ 2A2 + ω2K 2 , (6.2)
where
η is the transmittance of the incident radiation
A is the detecting/illumination area
γ is the coefficient corresponding to the loss of heat per unit area of the detector to its surround-
ings due to its increase in temperature
K is the illumination energy for obtaining 1°C temperature rise given by
K = ρcp Ah, (6.3)
where
ρ is the density of the pyroelectricmaterial
cp is the specific heat
h is the thickness of the detector (refer to Figure 6.3a)
Equation 6.2 indicates that the temperature oscillation is obtained from the energy balance between
the irradiation and heat dissipation.
The current responsivity, ri, is defined by
ri = (1/WA )(dq dt ). (6.4)
Since the charge generated by a temperature rise ΔT is given as
q = pAΔT , (6.5)

VS
V0
CD R CA RL
V=0
FIGURE 6.6 Amplifier for a pyroelectric IR detector.
using Equation 6.2, we obtain
( )
−1/2
ri = ηpωA γ 2A2 + ω2K 2 . (6.6)
Introducing a thermal time constant by the ratio of heat generation (absorption) K over heat radia-
tion (dissipation) γA as
τD = K γ A, (6.7)
we obtain finally
( )
−1 2
ri = ηpωγ −1 1 + ω2 τD2 . (6.8)
When ωτD >> 1, ri = (ηp/ρcph). In order to increase ri, neglecting the size or surface effect, the value
(p/ρcp) should be increased.
Figure 6.6 shows an amplifier circuit for measuring a pyroelectric voltage signal. The resistance
R is relatively high and is inserted to remove the charge after it is thermally induced on the pyro-
electric (CD). The transistor must have high impedance (e.g., FET).
Next, the voltage responsivity for such an amplifier is expressed as
rv = (1 WA )(dV dt ) = ri Z , (6.9)
where Z is the impedance of the detector–amplifier combination. Assuming RL << R,
( )
−1 2
Z = R 1 + ω2 τ 2E (6.10)
where τE = R(CD + CA), and CD and CA are the capacitances of the detector and the amplifier.
Therefore, Equation 6.9 may be written as
( ) (1 + ω τ )
−1 2 −1 2
rv = ηpωR γ −1 1 + ω2 τD2 2 2
E . (6.11)
At a high frequency (>>1/τD, 1/τE), we obtain
rv = ηp ρcpεAω , (6.12)
assuming that CD > CA. In order to increase rv, neglecting again size or surface effects, the value
(p/ρcpε) should increase. Note that rv differs from ri by a factor of (1/ε). The rv decreases with fre-
quency at high frequencies, but is relatively independent of frequency between 1/τD (0.1–10 Hz) and
1/τE (0.01 Hz).3 Thus, in practice, the irradiation chopping frequency is chosen just between 1/τD
and 1/τE.

6.1.3 FIGURES OF MERIT

A pyroelectric sensor is a device for transducing optical/thermal energy to electrical energy, and its
sensitivity or figure of merit (FOM) can be evaluated in several ways; for example, in terms of p,
p/cp, or p/(cpε). These are useful FOMs because the temperature change of the sample is larger for
the smaller specific heat (cp) material under constant heating, and the voltage generated by a certain
amount of pyrocharge becomes larger for the smaller permittivity (ε) material (refer Section 6.1.2).
Table 6.1 summarizes several FOMs. Table 6.2 lists the FOMs of several pyroelectric materials.4
EXAMPLE PROBLEM 6.2

Assuming a second-order phase transition for the Landau free energy
F (P,T ) = (1 2 )αP 2 + (1 4 )βP 4 , (P6.2.1)
α = (T − T0 ) ε 0 C , (P6.2.2)
formulate the temperature dependence of the following three FOMs for a pyroelectric detector:
p, (p/cp), and (p/cpε0ε). Here p is the pyroelectric coefficient, cp is the specific heat, and ε is the
relative permittivity.
TABLE 6.1
Figures of Merit for Pyroelectric Materials
FOM Application
p/cp Low-impedance amplifier
p/(cpε) High-impedance amplifier
p/(cpαε) Thermal imaging device (vidicon)
p/cp(ε tan δ)1/2 High-impedance amplifier when the pyroelectric
element is the main noise source
Note: p, pyroelectric coefficient; cp, specific heat; ε, absolute permittivity;

α, thermal diffusivity; tan δ, dielectric loss.
TABLE 6.2
Room-Temperature Properties of Various Pyroelectric Detector Materials and Some
FOMs for Their Detector Operation
p/cp (n A p/(cpe¢) p/(cpe≤)
Material p (n C/cm2/K) e¢/e0 cp (J/cm3/K) cm/W) (V/cm2/J) (cm3/J)1/2
TGS 30 50 1.7 17.8 4000 0.149
LiTaO3 19 46 3.19 6.0 1470 0.050
Sr1/2Ba1/2Nb2O6 60 400 2.34 25.6 720 0.030
PLZT(6/80/20) 76 1000 2.57 29.9 340 0.034
PVDF 3 11 2.4 1.3 1290 0.009

Solution
The polarization for zero applied field is obtained from
⎡⎣(T − T0 ) ε 0C ⎤⎦ PS + βPS3 = 0. (P6.2.3)
For a ferroelectric/pyroelectric phase, T < T0, the minimum of the Landau free energy is obtained at
PS = (T0 − T βε 0C ). (P6.2.4)
Since the dielectric constant ε is calculated as
(
1 ε = ε 0 (∂P ∂E ) = ε 0 α + 3β P 2 , ) (P6.2.5)
we obtain
ε = C ⎡⎣2 (T0 − T )⎤⎦ , (T < T0 ). (P6.2.6)
Concerning the specific heat, modulation around the phase transition range from the Debye spe-
cific heat cp0 (we consider this almost constant) is neglected (cp = cp0).
From the above relations, we can calculate the FOMs:
p = − (∂Ps ∂T ) = 1/2 (βε 0C ) (T0 − T )−1 2 ,

−1 2
(P6.2.7)
p cp = (1 2 )(βε 0C ) cp0 −1 (T0 − T )

−1 2 −1 2
, (P6.2.8)
−3 2
(ε0C ) (T0 − T ) ,
−1 2 −1 12
p cp ε 0 ε = β cp0 (P6.2.9)
Thus, p and (p/cp) μ (T0 − T)−1/2, while (p/cpε0ε) μ (T0 − T)1/2, meaning that the former FOMs
are large around the Curie temperature, while the last FOM is small around TC. Refer to Chapter
Problem 6.1 for the first-order phase transition. Actual experimental results for a barium titanate-
based sample (first-order phase transition) are shown in Figure 6.7.
Improvement of the sensing characteristics has been attempted by using composites of
pyroceramics and polymers.5 These composites have additional benefits, on top of the primary
Bias = 1 kV/cm
Figure of merit p/cpε
800
Figure of merit p/cpε
300
600 2 kV/cm
3 kV/cm 200
400 4 kV/cm
100
200
0
18 20 22 24 0 5 10
(a) Temperature (°C) (b) Bias field (kV/cm)
FIGURE 6.7 FOM (p/cpε) change with temperature (a) and bias field (b) for the Ba0.67Sr0.33TiO3-based
ceramic. The poling voltage is the same as the biasing voltage. (a) Note that the bias field stabilizes the tem-
perature characteristics significantly. (b) Maximum black body (490°C) response of a 50 μm thick sample
of BST at a chopper frequency of 40 Hz. (From Kulwicki, B.M. et al., Proc. Int’l Symp. Appl. Ferroelectr.,
Greenville, SC, IEEE, p. 1, 1992.)

pyroelectric effect, a secondary effect enhances the sensing capability. Stress due to thermal
expansion difference between the ceramic and polymer generates additional electric charge through
the piezoelectric effect, increasing sensitivity.
There are also examples of enhancements of FOMs by driving conditions. For instance, Texas
Instruments used a bias voltage on a pyroelectric (Ba,Sr)TiO3 ceramic during detection, and reported
a remarkable enhancement of the FOM p/cpε.6 Figure 6.7 shows the FOM (p/cpε) change with tem-
perature (a) and bias field (b) for the Ba0.67Sr0.33TiO3-based ceramic. The poling voltage is the same
as the biasing voltage. Note that the bias field stabilizes the temperature characteristics significantly
in (a). Measurement was done with maximum black body (490°C) response of a 50 μm thick sample
of BST at a chopper frequency of 40 Hz.6
6.2 TEMPERATURE/INFRARED SENSORS

Figure 6.8 shows a typical structure for a polymer pyroelectric IR sensor. A thin polyvinyl-
difluoride (PVDF) ferroelectric film is suspended in a metal ring to facilitate the temperature
change, and is encapsulated in a metal case with a silicon window. Note that silicon is transparent
to IR light. A FET amplifier is set very close to the film in order to amplify the small electric signal
generated in the film immediately after the absorption of IR light.
In practice, a pyrosensor requires an IR light (thermal ray) chopper, because the electrical signal
can only be detected at the transient stage of the light illumination and shutoff. An electromagnetic
(EM) motor is conventionally used to drive a light-chopper mechanism, but the EM motor is bulky
and has a low efficiency, and increases EM noise. A piezoelectric bimorph chopper was developed
by Kuwano et al.,7 which allows for miniaturization of the pyrosensors in a package with a light
chopper, also minimizing the EM noise (see Figure 6.9). Two optical slits were swung with a pair
of piezoelectric bimorphs.
6.3 INFRARED IMAGE SENSORS

Figure 6.10a illustrates the structure of a pyrovidicon tube, which visualizes a thermal-distribution
(i.e., IR) image.8 Light emitted from an object, which should be warmer than the surrounding temper-
ature, is filtered with a germanium lens cutting shorter wavelength than IR beam (long wavelength),
Cushion ring
Silicon window
Pyro film ring
ø 8.3
5.8
Supporter FET Ground ring

Substrate Reflector
FIGURE 6.8 A polymer-based (PVDF) pyroelectric IR sensor.

Infrared ray (incoming)

Case
Room-temperature
detector
Slit Piezoelectric
bimorph
Piezoelectric bimorph
Infrared detector Package
FIGURE 6.9 Swing-type pyroelectric temperature sensor (by Sanyo). (From Shibata, K. et al., Jpn. J. Appl.
Phys., 24, 181, 1985.)
Germanium Focus and

lens Target Mesh Wall anode scan coils
Grid
Cathode
Signal lead
Chopper
Electron beam
(a) Germanium window First anode
Target element capacitance Ct

Beam impedance
Signal plate
Cathode Z
i
V
Input
capacitance Ci
(b)
FIGURE 6.10 Structure of a pyrovidicon tube (a) and its equivalent circuit (b).
which is focused onto the pyroelectric target through an optical chopper. The temperature distribu-
tion of the object is represented on the target as a shrunk image, leading to a pyroelectric voltage
distribution. This is monitored from the back surface of the target by electron-beam scanning using
a conventional TV tube. Figure 6.10b shows the equivalent circuit for the pyrovidicon.
One of the disadvantages of the pyrovidicon is the degradation of the image over a long period
of usage due to thermal diffusion on the target. From this sense, the FOM p/(cpαε) is essential,
including the thermal diffusivity α. The thermal diffusivity can also be improved by modifying
the structure. Pedder et al. proposed a segmented target design to solve the diffusion problem.9
Figure 6.11 shows the microscopic structure of a D-TGS [deuterated triglycine sulphate, (ND2CD2

FIGURE 6.11 IR image target of a D-TGS crystal with divided fine segments (19 μm width, 16 μm depth,
and 25 μm pitch).
COOD)3D2SO4] target, in which TGS crystal target was divided into fine segments (19 μm width,
16 μm depth, and 25 μm pitch). Figure 6.12 shows two example still images taken by a pyrovidicon
on a dark night.
Figure 6.13a introduces one of the most recent thermal IR cameras. With this camera, Figure 6.13b
demonstrates heat generation monitoring on a PZT rectangular plate under its first resonance mode,
and Figure 6.13c is its corresponding thermograph on a numerical chart. We can observe significant
heat generation around the nodal line area on this resonator, where the maximum stress and strain
generate the heat via the elastic loss. The physical background is discussed in detail in Section 3.1.3.
6.4 ELECTROCALORIC DEVICES

6.4.1 ELECTROCALORIC EFFECT
The converse effect of pyroelectricity is “electrocaloric effect.” By applying the electric field on a
pyrocrystal, we can expect a temperature rise/fall according to the polarity. The cooling function is
of particular interest, aiming at the refrigeration application.
6.4.1.1 Phenomenological Formula

Let us review the thermodynamic phenomenological theory (Section 2.2.1) discussed basically
in the form of expansion series of the free energy as a function of the physical properties; one of
polarization P and electric field E, one of temperature T and entropy S. We take into account the
expansion series in terms of P and temperature:
F (P,T ) = a2 P 2 + a4 P 4 + a6 P 6 + … + b1T + b2T 2 + … + c1TP 2 + …
From S = − (∂F ∂T )P = −b1, and that a constant entropy is meaningless, we take b1 = 0. The term b2T 2
is a higher order term to be neglected. Thus we adopt only c1TP2. It is important to understand that
the product TP2 of the two parameters (P2 and T) explain the coupling effect; that is, T change causes

FIGURE 6.12 Images taken by a pyrovidicon on a dark night.

23.00 62.25
(a) (b)
70
60
Temperature (°C)
50
40
30
20
10
0 10 20 30 40
(c) Length (mm)
FIGURE 6.13 (a) Thermal IR camera, (b) heat generation monitoring on a PZT rectangular plate under its
first resonance mode, and (c) thermograph on a numerical chart.
P change to keep the same free energy (this effect is called “pyroelectric” effect), or E application
causes T change (this is called “electrocaloric” effect). By combining a2P2 and c1TP2, we introduce
⎛ T − T0 ⎞
(1 2 )α = a2 P 2 + c1TP 2 = (1 2 )⎜ ⎟
⎝ ε0C ⎠
We also introduce the following notations: (1/4)β = a4, (1/6)γ = a6, and we define the Landau free
energy as
F (P,T ) = (1 2 )α (T )P 2 + (1 4 )βP 4 + (1 6 )γP 6 , (6.13)
where α(T) = (T − T0)/ε0 C. Here, electromechanical coupling is tentatively neglected.

Now, let us consider the entropy S of the crystal as a function of temperature T and electric
field E:
dS = (cp T )dT + ( ∂S ∂E )dE, (6.14)
where cp is the specific heat. Note that T dS (=dQ) is the heat flow into this crystal. When the
crystal is set thermally isolated (i.e., adiabatic condition), the temperature change of the crystal is
expressed by an electric field change as

(dT )S = − (T cp )( ∂S ∂E )T (dE )S . (6.15)
By adopting a general Maxwell relation (∂S/∂E)T = (∂P/∂T)E, Equation 6.15 becomes
(dT )S = − (T cp )(∂P ∂ T )E (dE )S . (6.16)
Taking into account the pyroelectric coefficient p = −(∂P/∂T)E, we obtain finally
(dT )S = p (T cp )(dE )S . (6.17)
Equation 6.17 is an essential relation for considering the electrocaloric effect; in an adiabatic condi-
tion, temperature change occurs in this crystal that is directly proportional to the applied electric
field. In a regular ferroelectric material with p > 0 (PS decreases with T increase), total proportional
constant is positive, leading to cooling function with negative electric field on the crystal (E is in
opposite direction to PS). Higher p and smaller cp materials are preferred. Note also that the higher
T enhances the temperature change ΔT.
6.4.1.2 Electrocaloric Materials

Electrocaloric materials were the first focus on military applications during World War II, in partic-
ular, for cooling systems in submarines. A significant scientific interest in electrocaloric materials
followed in the 1960s and 1970s, but devices were not commercially exploited as the electrocaloric
effects were insufficient for practical applications.
In the beginning of twenty-first century, necessity for energy conservation accelerated the elec-
trocaloric material/device development for cooling systems. The triggering work done by the Penn
State University group (S. Priya and K. Uchino) initially verified large pyroelectric/electrocaloric
effects in PZN–PT single crystals, as shown in Figure 6.14.10 Electrocaloric effect was measured
in <111>-oriented 0.92Pb(Zn1/3Nb2/3)O3–0.08PbTiO3 single crystals as a function of temperature
0.25
<111>
0.20
0.15
ΔT (°C)
0.10
0.05
0.00
40 80 120 160 200

Temperature (°C)
FIGURE 6.14 Electrocaloric effect in <111>-oriented 0.92Pb(Zn1/3Nb2/3)O3–0.08PbTiO3 single crystals as

a function of temperature under an electric field of 1 kV/mm. (From Priya, S. and Uchino, K., Jpn. J. Appl.
Phys., 42, 5158, 2003. With permission.)

under an electric field of 1 kV/mm, which exhibited reasonably large temperature decrease in a bulk
samples. This research motivated various researchers and device prototypes.
A. S. Mischenko et al. reported that thin films of PZT showed the strongest electrocaloric response
yet reported, with the materials cooling down by up to 12°C at 25 V.11 Figure 6.15a shows the tem-
perature changes ΔT due to applied electric field ΔE. By fixing E2 = 776 kV/cm, ΔE = E2 − E1 was
changed. The peak value of ΔT = 12 K occurred in ΔE = 480 kV/cm at T = 226°C, where p = |∂PS/∂T|
is maximized (see the inset figure). They demonstrated 0.48 K/V electrocaloric effect in 350 nm thick
PbZr0.95Ti0.05O3 films near the Curie temperature of 222°C (Figure 6.15b). Q. Zhang et al. at the
Pennsylvania State University developed a ferroelectric polymer that can achieve 12°C of cooling.12
ΔE (kV/cm)
12 E (kV cm–1)
70 776
P (μC cm–2)
480
60 429
50 295
10
40
30 380
20
8 0 100 200 300
T(°C)
ΔT (°C)
280
6
4 180
2 80
30
0
50 100 150 200 250 300
(a) T (°C)
(b)
FIGURE 6.15 (a) Temperature changes ΔT due to applied electric field ΔE. By fixing E2 = 776 kV/cm, ΔE =
E2 − E1 was changed. The peak value of ΔT = 12 K occurred in ΔE = 480 kV/cm at T = 226°C, where p =
|∂PS/∂T| is maximized (see the Inset figure). (b) Thin film-type electrocaloric demo kit. (From Mischenko,
A.S. et al., Science, 311, 1270, 2006. With permission.)

Chapter Essentials
1. Merits of pyrosensors compared to other IR-sensor materials such as semiconductors:

(a) Wide range of response frequency.
(b) Use at room temperature.
(c) Quick response in comparison with other temperature sensors.
(d) High-quality (optical-grade homogeneity, etc.) materials are NOT necessary for
pyrosensors.
2. Figures of Merit for pyroelectric materials:
FOM Application
p/cp Low-impedance amplifier
p/(cpε) High-impedance amplifier
p/(cpαε) Thermal imaging device (vidicon)
p/cp(ε tan δ)
1/2 High-impedance amplifier when the pyroelectric element
is the main noise source
Note: p, pyroelectric coefficient; cp, specific heat; ε, absolute permittivity; α, ther-
mal diffusivity; tan δ, dielectric loss.
3. Thick film structure is essential for quick responsivity, and a light-chopper mechanism
(e.g., piezoelectric bimorphs) is the key to miniaturization.
4. Electrocaloric effect under an adiabatic condition
ΔT = p (T cp )ΔE
where
p is the pyroelectric coefficient (p = |∂PS/∂T|)
cp is the specific heat
Check Point
1. (T/F) The Curie temperature of a pyroelectric sensor material should be 200°C higher than
room temperature. True or False?
2. Provide the FOM for pyroelectric materials for thermal imaging device applications. Use the nota-
tions: p, pyroelectric coefficient; cp, specific heat; ε0ε, permittivity; and α, thermal diffusivity.
3. (T/F) When the IR beam chopping frequency is reasonably high (such as 100 Hz) the cur-
rent responsivity is frequency-independent. True or False?
4. (T/F) When the IR beam chopping frequency is reasonably high (such as 100 Hz) the cur-
rent responsivity is directly proportional to the pyromaterial film thickness. True or False?
5. Though the pyroelectric coefficient p is not large, PVDF is popularly used for a pyrosensor.
What is the key reason?
Chapter Problems
6.1 Assuming the first-order phase transition for the Landau free energy, calculate the tem-
perature dependence of the FOMs for a pyroelectric detector: p, p/cp, and p/cpε0ε.
6.2 There is a PLZT (6/80/20) ceramic disk with 1 cm2 in area and 100 μm in thickness elec-
trically poled along the thickness with transparent electrodes. When the sample is illu-
minated with a laser light (power: 10 mW/cm2) for 0.1 s, calculate the following values:
(a) The temperature rise of the sample
(b) The charge generated on the surface transparent electrode
(c) The open-circuit voltage generated

Sharp phase transition
Spontaneous polarization
(a)
Diffuse phase transition
(b)
Successive phase transition
(c) RT Temperature
FIGURE 6.16 Spontaneous polarization vs. temperature relations: (a) sharp phase transition, (b) diffuse
phase transition, and (c) successive phase transition materials.
Assume that all the light energy passes the electrode, and is absorbed by the sample,
and that neither heat loss nor electric loss is taken into account. Refer to Table 6.2 for
the necessary data.
Hint
Total heat energy: 10 (mW/cm2) × 1 (cm2) × 0.1 (s) = 1 (mJ)
Sample volume v: 1 (cm2) × 0.01 (cm) = 0.01 (cm3)
Temperature rise ΔT: 1 (mJ)/[2.57 (J/cm3/K1) × 0.01 (cm3)] = 0.039 (K)
6.3 Consider three materials: (a) sharp phase transition, (b) diffuse phase transition, and (c)
successive phase transition materials, respectively, with the spontaneous polarization vs.
temperature relations as illustrated in Figure 6.16a through c. Discuss the merits and demer-
its of each from a pyrodetector application viewpoint with respect to the following:
(1) Magnitude of p
(2) Relative permittivity
(3) Temperature stability
(4) Aging
REFERENCES
1. A. Rogalski: Sensor. Mater., 12(5), 233 (2000).
2. J. M. Herbert: Ferroelectric Transducers and Sensors, p. 267, Gordon & Breach, New York (1982).
3. S. G. Porter: Pyroelectricity and Its Use in Infrared Detectors, Plessey Optoelectronics and Microwave
Ltd., Towcester, U.K. (1980).
4. M. E. Lines and A. M. Glass: Principles and Applications of Ferroelectrics and Related Materials,
Clarendon Press, Oxford (1977).
5. A. S. Bhalla, R. E. Newnham, L. E. Cross, W. A. Schulze, J. P. Dougherty, and W. A. Smith: Ferroelectrics,
33, 139 (1981).
6. B. M. Kulwicki, A. Amin, H. R. Beratan, and C. M. Hanson: Proc. Int’l Symp. Appl. Ferroelectr.,
Greenville, SC, IEEE, p. 1 (1992).
7. K. Shibata, K. Takeuchi, T. Tanaka, S. Yokoo, S. Nakano, and Y. Kuwano: Jpn. J. Appl. Phys., 24(Suppl.
24-3), 181 (1985).
8. R. G. F. Taylor and H. A. H. Boot: Contemp. Phys., 14, 55 (1973).
9. D. J. Warner, D. J. Pedder, I. S. Moody, and J. Burrage: Ferroelectrics, 33, 249 (1981).
10. S. Priya and K. Uchino: Jpn. J. Appl. Phys., 42, 5158–5164 (2003).
11. A. S. Mischenko, Q. Zhang, J. F. Scott, R. W. Whatmore, and N. D. Mathur: Science 311, 1270 (2006).
12. Science 8 August (2008).

7 Piezoelectric Devices
Certain materials produce electric charges on their surfaces when mechanical stress is applied. The
induced charges are proportional to the mechanical stress. This is called the direct piezoelectric
effect and was discovered in quartz by Pierre and Jacques Curie in 1880. Materials showing this
phenomenon also conversely have a geometric strain proportional to an applied electric field. This
is the converse piezoelectric effect, discovered by Gabriel Lippmann in 1881. The root of the word
“piezo” means “pressure” in Greek; hence the original meaning of the word piezoelectricity implied
“pressure electricity.”
Piezoelectricity is extensively utilized in the fabrication of various devices such as transducers,
sensors, actuators, surface acoustic wave (SAW) devices, frequency control, and so on. In this chap-
ter we describe the piezoelectric materials that are used, and then the various applications of these
piezoelectrics.1–4
7.1 PIEZOELECTRIC MATERIALS AND PROPERTIES

7.1.1 PIEZOELECTRIC FIGURES OF MERIT
There are five important figures of merit in piezoelectrics: the piezoelectric strain constant d, the
piezoelectric voltage constant g, the electromechanical coupling factor k, the mechanical quality
factor Qm, and the acoustic impedance Z. These figures of merit are considered in this section.
Piezoelectric strain constant d
The magnitude of the induced strain x by an external electric field E is represented by this figure of
merit (an important figure of merit for actuator applications):
x = dE. (7.1)
Piezoelectric voltage constant g

The induced electric field E is related to an external stress X through the piezoelectric voltage con-
stant g (an important figure of merit for sensor applications):
E = gX . (7.2)
Taking into account the relation, P = dX, we obtain an important relation between g and d:
g = d /ε0ε (ε: relative permittivity). (7.3)
EXAMPLE PROBLEM 7.1

Obtain the relationship between the piezoelectric d and g constants, which indicates the strain
per unit electric field and the electric field per unit stress.
Solution
From the fundamental piezoelectric equations
x = s E X + dE, (P7.1.1)
161

P = dX + ε0 ε X E, (P7.1.2)
the actuator figure of merit d (external X = 0) is given by Equation P7.1.1: x = dE, and the sensor
figure of merit d (external E = 0) is given by Equation P7.1.2: P = dX. Here, P is polarization and
sE is elastic compliance under constant electric field. The polarization P induced in a material
with ε0ε results in an electric field (under an open-circuit condition) of
E = P ε0ε
X
(
= dX ε 0 ε X . ) (P7.1.3)
Therefore, taking into account E = gX, we can derive
g = d/ε 0 ε X . (P7.1.4)
Electromechanical coupling factor k

The terms, electromechanical coupling factor, energy transmission coefficient, and efficiency are
sometimes confused.5 All are related to the conversion rate between electrical energy and mechani-
cal energy, but their definitions are different.6
a. The electromechanical coupling factor k
k 2 = (Stored mechanical energy/Input electrical energy) (7.4)
or
k 2 = (Stored electrical energy/Input mechanical energy) (7.5)
Let us calculate Equation 7.4, when an electric field E is applied to a piezoelectric material. See
Figure 7.1a left. Since the input electrical energy is (1/2) ε0εE2 per unit volume and the stored
mechanical energy per unit volume under zero external stress is given by (1/2)x2/s = (1/2)
(dE)2/s, k2 can be calculated as
k 2 = [(1/2)(dE )2 /s]/[(1/2)ε0εE 2 ]
= d 2 /ε0εs. (7.6)
b. The energy transmission coefficient l max

Not all the stored energy can be actually used, and the actual work done depends on the
mechanical load. With zero mechanical load or a complete clamp (no strain), no output
work is done. The energy transmission coefficient is defined by
λ max = (Output mechanical energy/Input electrical energy)max (7.7)
or equivalently
λ max = (Output electrical energy/Input mechanical energy)max (7.8)
The difference between Equations 7.4 and 7.5 with Equations 7.7 and 7.8 is whether the energy is
“stored” or “spent.”
Let us consider the case where an electric field E is applied to a piezoelectric under constant exter-
nal stress X (<0, because a compressive stress is necessary to work to the outside). This corresponds

Piezoelectric Devices 163
x
Mass
E dE
x
x
E
Ps Ps
0 E
Piezo actuator
sX
(a) (b)
x P Input
Output energy
dE
energy
ε0ε E + dX
dE + sX
–dE/s E
X 0 X 0 E
dX
sX
(c) (d)
FIGURE 7.1 Calculation of the input electrical and output mechanical energy: (a) load mass model for the
calculation, (b) electric field vs. induced strain curve, (c) stress vs. strain curve, and (d) electric field vs. polar-
ization curve.
to the situation that a mass is put suddenly on the actuator, as shown in Figure 7.1a. Figure 7.1b shows
two electric field versus induced-strain curves, corresponding to two conditions; under the mass load
and no mass. Because the area on the field-strain domain does not mean the energy, we should use the
stress–strain and field-polarization domains in order to discuss the mechanical and electrical energy,
respectively. Figure 7.1c illustrates how to calculate the mechanical energy. Note that the mass shrinks
the actuator initially by sX (s: piezomaterial’s compliance and X < 0). The mechanical energy sX2 is
a sort of “loan” of the actuator credited from the mass, which should be subtracted later. This energy
corresponds to the hatched area in Figure 7.1c. By applying the step electric field, the actuator expands
by the strain level dE under a constant stress condition. This is the mechanical energy provided from
the actuator to the mass, which corresponds to |dEX|. Like paying back the initial “loan,” the output
work (from the actuator to the mass) can be calculated as the area subtraction (shown by the dotted
area in Figure 7.1c):
∫ (− X )dx = − (dE + sX )X . (7.9)
Figure 7.1d illustrates how to calculate the electrical energy. The mass load X generates the “loan”
electrical energy by inducing P = dX (see the hatched area in Figure 7.1d). By applying a sudden
electric field E, the actuator (like a capacitor) receives the electrical energy of ε0εE2. Thus, the total
energy is given by the area subtraction (shown by the dotted area in Figure 7.1d):
∫ (E )dP = (ε ε E + dX )E.
0 (7.10)
We need to choose a proper load to maximize the energy transmission coefficient. From the maxi-
mum condition of
λ = − (dE + sX ) X/(ε 0 εE + dX )E, (7.11)

we can obtain
−2
(1 k ) − 1 ⎥⎦⎤ (1 k ) − 1 ⎤⎥⎦
2
λ max = ⎡(1 k ) − 2
= ⎡(1 k ) + 2
. (7.12)
⎢⎣ ⎢⎣
Refer to Example Problem 7.2. Notice that
k 2 /4 < λ max < k 2 /2, (7.13)
depending on the k value. For a small k, λmax = k2/4, and for a large k, λmax = k2/2.
It is also worth noting that the maximum condition stated above does not agree with the condi-
tion which provides the maximum output mechanical energy. The maximum output energy can
be obtained when the dotted area in Figure 7.1c becomes maximum under the constraint of the
rectangular corner point tracing on the line (from dE on the vertical axis to −dE/s on the horizontal
axis). Therefore, the load should be a half of the maximum generative stress and the mechanical
energy: −[dE − s(dE/2s)](−dE/2s) = (dE)2/4s. In this case, since the input electrical energy is given by
[ε0εE + d (−dE/2s)]E
λ = 1/2[(2/k 2 ) − 1], (7.14)
which is close to the value λmax when k is small, but has a different value when k is large, that is
predicted theoretically.
c. The efficiency η
η = (Output mechanical energy)/(Consumed electrical energy) (7.15)
or
η = (Output electrical energy)/(Consumed mechanical energy). (7.16)
The difference of the efficiency definition from Equations 7.7 and 7.8 is “input” energy and “con-
sumed” energy in the denominators. In a work cycle (e.g., an electric field cycle), the input electri-
cal energy is transformed partially into mechanical energy and the remaining is stored as electrical
energy (electrostatic energy like a capacitor) in an actuator. In this way, the ineffective electrostatic
energy can be returned to the power source, leading to near 100% efficiency, if the loss is small.
Typical values of dielectric loss in lead zirconate titanate (PZT) are about 1%–3%.
EXAMPLE PROBLEM 7.2

Prove that the maximum energy transmission coefficient λmax in a piezoelectric actuator is
expressed as follows using its electromechanical coupling factor k:
(1 k )− 1 ⎤⎥⎦ .
2
λ max = ⎡(1 k )− 2
(P7.2.1)
⎢⎣
Solution
The energy transmission coefficient is defined by
λ = (Output mechanical energy/Input electrical energy).

Considering the case where an electric field E is applied to a piezoelectric under a constant exter-
nal stress X, λ can be calculated as
λ = − xX PE
= − (dE + sX )X (ε 0 ε E + dX )E
= − ⎡⎢ d (X E )+ s (X E ) ⎤⎥ ⎡⎣ ε 0 ε + d (X E )⎤⎦ .
2
(P7.2.2)
⎣ ⎦
We need to determine an appropriate stress X under a certain applied field E so as to maximize

the λ value. Letting y = X/E, then
λ = − (sy 2 + dy)/(dy + ε 0 ε). (P7.2.3)
The maximum λ can be obtained when y satisfies
(dλ dy) = ⎡⎣− (2sy + d )(dy + ε 0 ε ) + (sy2 + dy)d⎤⎦ (dy + ε 0 ε )

2
= 0. (P7.2.4)
Then, from y 02 + 2(ε0ε/d)y0 + (ε0ε/s) = 0
y0 = (ε 0 ε d )⎡ −1 +
⎢⎣ (1 − k )⎤⎥⎦ .
2
(P7.2.5)
( )
Here, k2 = d 2/(sε0ε). Note that only y0 =(ε 0 ε /d ) ⎡⎢ −1 + 1 − k 2 ⎤⎥ is valid for realizing the mean-
⎣ ⎦
⎡
ingful maximum point, since y0 = (ε 0 ε / d ) −1 − 1 − k
⎣⎢
(
2 ⎤
⎦⎥
)
and y0 = (ε 0 ε/d ) ⎡−1+ 1 − k 2 ⎤
⎣⎢
(
⎦⎥
)
provided (d2λ/dy2) > 0 and < 0, respectively.
By putting y = y0 into λ(y), we can get the maximum value of λ:
λ max = −s ⎡⎣ −2 (ε 0 ε d )y 0 − (ε 0 ε s)⎤⎦ + dy0 ) (dy0 + ε 0 ε )
( )
= ⎡ dy0 2 k 2 − 1 + ε 0 ε ⎤ (dy0 + ε 0 ε )
⎣ ⎦
⎧
= ⎨ ⎡ −1 +
⎩ ⎢⎣
(1 − k )⎥⎦⎤ (2 k
2 2
) ⎫ ⎧
− 1 + 1⎬ ⎨ ⎡ −1 +
⎭ ⎩ ⎢⎣
(1 − k )⎥⎦⎤ + 1⎭⎬⎫ .
2
2
⎡ ⎛ 1⎞ ⎤
= ⎢(1 k )− ⎜ 2 ⎟ − 1 ⎥ . (P7.2.6)
⎢⎣ ⎝k ⎠ ⎥⎦
An elastic material with strain x under stress X stores mechanical energy (1/2)xX per unit volume.
Similarly, a dielectric material with polarization P under electric field E stores electrical energy
(1/2)PE per unit volume. However, in Figure 7.1a model, the mechanical and electrical energies
were calculated by xX and PE, respectively, without putting (1/2). Why?
Answer
When the stress is gradually applied to the material, the strain also is induced gradually, following
the Hook’s Law: x = sX. Thus, the total mechanical energy is calculated by

UM = ∫ x dX = ∫ sX dX = (1 2)sX 2 = (1 2)xX .
(Q7.1.1)
To the contrary, when the constant stress is applied suddenly, e.g., a mass is put on the actuator,
as in this case, the mechanical energy is calculated by
UM = ∫ X dx = X ⋅ ∫ dx = X ⋅ x. (Q7.1.2)
Similarly, when the electric field is gradually applied to the material, the polarization is induced in
proportion to E, and the total electrical energy is
UE = ∫ P dE = ∫ ε 0 εE dE = (1 2)ε 0 εE = (1 2)PE. (Q7.1.3)

2
When the step field is applied, the energy is
UE = ∫ E dP = E ⋅ ∫ dP = EP. (Q7.1.4)
Mechanical quality factor Q M

The mechanical quality factor, Q M, is a parameter that characterizes the sharpness of the electro-
mechanical resonance spectrum. When the motional admittance Ym is plotted around the resonance
frequency ω0, the mechanical quality factor Q M is defined with respect to the full width [2Δω] at
Ym / 2 as
QM =ω0 /2Δω. (7.17)
Also note that Q M−1 is equal to the mechanical loss (tan δm). When we define a complex elastic com-
pliance, sE = sE′ – jsE″, the mechanical loss tangent is provided by tan δm = sE″/sE′. The Q M value is
very important in evaluating the magnitude of the resonant displacement and strain. The vibration
amplitude at an off-resonance frequency (dEL, L: length of the sample) is amplified by a factor
proportional to Q M at the resonance frequency. For example, a longitudinally vibrating rectangular
plate through the transverse piezoelectric effect d31 generates the maximum displacement given by
(8/π2)Q M d31EL. Refer to Frequently Asked Question 3.1.
Acoustic impedance Z
The acoustic impedance Z is a parameter used for evaluating the acoustic energy transfer between
two materials. It is defined, in general, by
Z 2 = (pressure /volume velocity). (7.18)

In a solid material
Z = ρc , (7.19)
where
ρ is the density
c is the elastic stiffness of the material
In more advanced discussions, there are three kinds of impedances; specific acoustic impedance
(pressure/particle speed), acoustic impedance (pressure/volume speed), and radiation impedance
(force/speed). See Ref. [6] for the details.
Why is acoustic impedance (or mechanical impedance) matching necessary for the transfer of
mechanical energy from one material to the other? Explain the reason conceptually.

Easy !
Work W = F × ΔL
F=0
“Pushing a curtain, and

Wow, pushing a wall”
heavy ! Japanese proverb,
meaning “Useless Task”
ΔL = 0
FIGURE 7.2 Concept of mechanical impedance matching.
Solution
The mechanical work done by one material on the other is evaluated by the product of the applied
force F and the displacement ΔL:
W = F × ΔL.
(Q7.2.1)
Figure 7.2 shows a conceptual cartoon illustrating two extreme cases. If the material is very soft,
the force F can be very small, leading to very small W (practically no work). This corresponds
to “pushing a curtain,” exemplified by the case when the acoustic wave is generated in water
directly by a hard PZT transducer. Most of the acoustic energy generated in the PZT is reflected
at the interface, and only a small portion of acoustic energy is transferred into water. On the other
hand, if the material is very hard, the displacement will be very small, again leading to very small
W. This corresponds to “pushing a wall.” Polymer piezoelectric polyvinylidene difluoride (PVDF)
cannot drive a hard steel part efficiently. Therefore, the acoustic impedance must be adjusted to
maximize the output mechanical power:
ρ1c1 = ρ2c2 ,
(Q7.2.2)
where
ρ is the density
c is the elastic stiffness
The subscripts 1 and 2 denote the two adjacent materials
In practice, acoustic impedance matching layers (Elastically intermediate materials between PZT
and water, such as a polymer. More precisely the acoustic impedance Z should be Z1 . Z2 ) are
fabricated on the PZT transducer to optimize the transfer of mechanical energy to water.
7.1.2 PIEZOELECTRIC MATERIALS

This section summarizes the current status of piezoelectric materials: single-crystal materials,
piezoceramics, piezopolymers, composites, and piezofilms. Table 7.1 shows the material parameters
of these piezoelectric materials.7,8 Quartz with the highest mechanical quality factor is used for low-
loss transducers. PZT family shows high d and k suitable for high-power transducers. Sm-doped
lead titanates exhibit extremely high mechanical coupling anisotropy kt/kp, suitable for medical
transducers. Piezopolymer PVDF has small permittivity, leading to high piezo g constant, in addi-
tion to mechanical flexibility, suitable to pressure/stress sensor applications.

TABLE 7.1
Piezoelectric Properties of Representative Piezoelectric Materials
Parameter Quartz BaTiO3 PZT 4 PZT 5 H (Pb,Sm)TiO3 PVDF-TrFE
d33 (pC/N) 2.3 190 289 593 65 33
g33 (10−3Vm/N) 57.8 12.6 26.1 19.7 42 380
kt 0.09 0.38 0.51 0.50 0.50 0.30
kp 0.33 0.58 0.65 0.03
ε3X/ε0 5 1700 1300 3400 175 6
QM >105 500 65 900 3–10
TC (°C) 120 328 193 355
Sources: Ito, Y. and Uchino, K., Wiley Encyclopedia of Electrical and Electronics Engineering, John Wiley
& Sons, New York, 1999; Smith, W.A., Proc. SPIE—The Int’l Society for Optical Engineering,
1733, 1992.
7.1.2.1 Single Crystals

Although piezoelectric ceramics are widely used for a large number of applications, single crystal
materials retain their utility, being essential for applications such as frequency-stabilized oscillators
and surface acoustic devices. The most popular single-crystal piezoelectric materials are quartz,
lithium niobate (LiNbO3), and lithium tantalate (LiTaO3). Single crystals are anisotropic, exhibit-
ing different material properties depending on the cut of the materials and the direction of bulk or
surface wave propagation.
Quartz is a well-known piezoelectric single crystal material. α-quartz belongs to the triclinic
crystal system with point group 32 and has a phase transition at 537°C to its β-form which is not
piezoelectric. Quartz has a cut with a zero temperature coefficient. For instance, quartz oscillators,
operated in the thickness shear mode of the AT-cut, are used extensively for clock sources in com-
puters, frequency-stabilized ones in TVs and VCRs. On the other hand, an ST-cut quartz substrate
with X-propagation has a zero temperature coefficient for SAW, and so is used for SAW devices with
high-stabilized frequencies. Another distinguished characteristic of quartz is an extremely high
mechanical quality factor Q M > 105.
Lithium niobate and lithium tantalate belong to an isomorphous crystal system and are composed
of oxygen octahedrons. The Curie temperatures of LiNbO3 and LiTaO3 are 1210°C and 660°C,
respectively. The crystal symmetry of the ferroelectric phase of these single crystals is 3m, and
the polarization direction is along c-axis. These materials have high electromechanical coupling
coefficients for SAW. In addition, large single crystals can easily be obtained from their melt using
the conventional Czochralski technique. Thus, both materials (LiNbO3 and LiTaO3) occupy very
important positions in the SAW device application field.
Single crystals of Pb(Mg1/3Nb2/3)O3 (PMN), Pb(Zn1/3Nb2/3)O3 (PZN) and their binary systems
with PbTiO3 (PMN–PT and PZN–PT) with extremely large electromechanical coupling factors are
discussed in the following section.
7.1.2.2 Polycrystalline Materials

Barium titanate BaTiO3 is one of the most thoroughly studied and most widely used ferroelectric
materials, after being discovered independently in Japan, United States, and Russia, during World
War II. Just below the Curie temperature (120°C), the vector of the spontaneous polarization points
in the [001] direction (tetragonal phase), below 5°C it reorients in the [011] (orthrhombic phase), and
below −90°C in the [111] direction (rhombohedral phase). The dielectric and piezoelectric proper-
ties of ferroelectric ceramic BaTiO3 can be affected by its own stoichiometry, microstructure, and

by dopants entering onto the A or B site in solid solution. Modified BaTiO3 ceramics with dopants
such as Pb or Ca ions have been developed to stabilize the tetragonal phase over a wider temperature
range and are used as commercial piezoelectric materials. The initial application was for Langevin-
type piezoelectric vibrators.
Piezoelectric Pb(Ti,Zr)O3 solid solutions (PZT) ceramics discovered in the United States in 1950s
have been widely used because of their superior piezoelectric properties. The phase diagram for the
PZT system (PbZrxTi1−xO3) is shown in Figure 7.3. The crystalline symmetry of this solid-solution
system is determined by the Zr content. Lead titanate also has a tetragonal ferroelectric phase
of perovskite structure. With increasing Zr content, x, the tetragonal distortion decreases and at
x > 0.52 the structure changes from the tetragonal 4mm phase to another ferroelectric phase of
rhombohedral 3m symmetry. The line dividing these two phases is called the morphotropic phase
boundary (MPB). The boundary composition is considered to have both tetragonal and rhombohe-
dral phases coexisting. Figure 7.4 shows the dependence of several piezoelectric d constants on com-
position near the MPB. The d constants have their highest values near the MPB. This enhancement
500
Cubic
400 a
a
a
Temperature (ºC)
300
Tetragonal Morphotropic Rhombohedral

200 phase
boundary a
PS c
a PS
100 a a
a
0
0 10 20 30 40 50 60 70 80 90 100
PbTiO3 PbZrO3
Mole % PbZrO3
FIGURE 7.3 Phase diagram of PZT.
800
600
d15
dij (×10–12 C/N)
400
d33
200
–d31
0
48 50 52 54 56 58 60
Mole % PbZrO3
FIGURE 7.4 Dependence of several d constants on composition near the MPB in the PZT system.

in piezoelectric effect is attributed to the increased ease of reorientation of the polarization under
an applied electric field. The theoretical background is illustrated in Frequently Asked Question 3.1
and Figure 3.6.
Doping the PZT material with donor or acceptor ions changes its properties dramatically. Donor
doping with ions such as Nb5+ or Ta5+ provides “soft” PZTs, like PZT-5, because of the easing of
domain motion due to the resulting Pb-vacancies. On the other hand, acceptor doping with Fe3+ or
Sc3+ leads to “hard” PZTs, such as PZT-8, because the oxygen vacancies will pin the domain wall
motion. Refer to Section 3.1.3 for the detailed description.
Since the PZT was protected by the patent of Clevite, United States, subsequently, PZT in ternary
solid solution with other perovskite phase has been investigated intensively by Japanese ceramic
companies. Examples of these ternary compositions are: PZTs in solid solution with Pb(Mg1/3Nb2/3)
O3 (Panasonic), Pb(Zn1/3Nb2/3)O3 (Toshiba), Pb(Mn1/3Sb2/3)O3, Pb(Co1/3Nb2/3)O3, Pb(Mn1/3Nb2/3)O3,
Pb(Ni1/3Nb2/3)O3 (NEC), Pb(Sb1/2Sn1/2)O3, Pb(Co1/2W1/2)O3, Pb(Mg1/2W1/2)O3 (Du Pont), all of which
were patented by different companies (almost all composition patents have already expired).
The end member of PZT, lead titanate has a large crystal distortion. PbTiO3 has a tetragonal
structure at room temperature with its tetragonality c/a = 1.063. The Curie temperature is 490°C.
Densely sintered PbTiO3 ceramics cannot be obtained easily, because they break up into a powder
when cooled through the Curie temperature as a result of the large spontaneous strain. Lead titanate
ceramics modified by adding a small amount of additives exhibit a high piezoelectric anisotropy.
Either (Pb,Sm)TiO39 or (Pb,Ca)TiO310 exhibits an extremely low planar coupling, that is, a large kt/kp
ratio. Here, kt and kp are thickness-extensional and planar electromechanical coupling factors, respec-
tively. Since these transducers can generate purely longitudinal waves through kt associated with no
transverse waves through k31, clear ultrasonic imaging is expected without the appearance of “ghost”
caused by the transverse wave. (Pb,Nd)(Ti,Mn,In)O3 ceramics with a zero temperature coefficient of
SAW delay have been developed as superior substrate materials for SAW device applications.11
EXAMPLE PROBLEM 7.3

Table 7.2 summarizes piezoelectric, dielectric, and elastic properties of typical PZTs: soft
PZT-5H, semihard PZT-4, and hard PZT-8. Using these data, answer to the following questions,
and learn the interrelations between these parameters.
a. From the values of piezoelectric d constant and permittivity ε, calculate the piezoelectric
g constants, and compare these calculations with the values provided in Table 7.2.
b. From the values of piezoelectric d constant, permittivity ε, and elastic compliance sE,
calculate the electromechanical coupling factors k, and compare these calculations with
the values provided in Table 7.2.
E /s E |, and piezoelectric Poisson’s ratio, |d /d | for
c. Calculate the elastic Poisson’s ratio, |s13 33 31 33
the above three PZTs, then, compare the similarity.
d. From the Q M values, calculate the elastic loss tan φ for these PZTs. Then, compare with
the dielectric loss tan δ for these three PZTs.
e. We apply 100 W electric energy on a k33-type PZT-4 rod. How much of electric energy
convert to mechanical energy stored in the PZT rod? Among that stored mechanical
energy, how much can we spend maximum for the outside work? In this procedure, how
much energy will be lost as heat?
Solution
a. The example calculation is made for PZT-4 with a k33-type rod:
g33 = d33 / ε 0 ε 33
E
= 285 × 10 −12 /8.854 × 10 −12 × 1300 = 24.8 × 10 −3 [Vm/N].
Calculate similarly for g31 and g15, and for other PZTs.

TABLE 7.2
Piezoelectric, Dielectric, and Elastic Properties of Typical PZTs
Soft PZT-5 H Semihard PZT-4 Hard PZT-8
EM coupling factor
kp 0.65 0.58 0.51
k31 0.39 0.33 0.30
k33 0.75 0.70 0.64
k15 0.68 0.71 0.55
Piezoelectric coefficient
d31 (10−12 m/V) −274 −122 −97
d33 593 285 225
d15 741 495 330
g31 (10−3 Vm/N) −9.1 −10.6 −11.0
g33 19.7 24.9 25.4
g15 26.8 38.0 28.9
Permittivity
ε33
X /ε
0
3400 1300 1000
ε11
X /ε
0
3130 1475 1290
Dielectric loss (tan δ) (%) 2.00 0.40 0.40
Elastic compliance
E (10−12m2/N)
s11 16.4 12.2 11.5
E
s12 −4.7 −4.1 −3.7
E
s13 −7.2 −5.3 −4.8
E
s33 20.8 15.2 13.5
E
s44 43.5 38.5 32.3
Mechanical QM 65 500 1000
Density ρ (103 kg/m3) 7.5 7.5 7.6
Curie temperature (°C) 193 325 300
b. The example calculation is made for PZT-4 with a k33-type rod:
k33 = d33 E
s33 ε 0 ε 33
X
= 285 × 10 −12 15.2 × 10 −12 × 8.854 × 10 −12 × 1300
= 0.68.
Calculate similarly for k31 and k15, and for other PZTs.
c. PZT-5H: |s13
E /sE | = 7.2 × 10 −12/20.8 × 10 −12 = 0.35
33
⎜d31/d33 ⎜ = 274 × 10 −12 /593 ×10 −12 = 0.46.
E /sE | = 5.3 × 10 −12/15.2 × 10 −12 = 0.35

PZT-4: |s13 33
⎜d31 /d33 ⎜ = 122 × 10 −12 /285 × 10 −12 = 0.42.
E /sE | = 4.8 × 10 −12/13.5 × 10 −12 = 0.36

PZT-8: |s13 33
⎜d31 /d33 ⎜ = 97 × 10 −12 /225 × 10 −12 = 0.43.

Input electrical energy 100

Usable
energy
Mechanical energy Loss 2
converted 50 Electrical energy
stored in a
capacitor 48 Dissipated
as heat
FIGURE 7.5 Energy conversion rate in a typical piezoelectric ceramic (PZT-4). From the input electrical
energy of 100 W, 50 W will be converted to mechanical energy stored in the piezocomponent. The remain-
ing energy is stored as electrical energy in the piezocapacitor. The loss (dissipated as heat) is 1% ∼ 2%.
The mechanical work from this actuator to the outside object is limited to only 1/3 ∼ 1/4 of the stored mechanical
energy.
Both Poisson’s ratios are close in the number, but |d31/d33| seems to be a little larger than
|s13
E /sE |.
33
d. PZT-5H: tan φ = 1/Q M = 1/65 = 0.015 « tan δ = 0.02.
PZT-4: tan φ = 1/Q M = 1/500 = 0.002 « tan δ = 0.004.
PZT-5H: tan φ = 1/Q M = 1/1000 = 0.001 « tan δ = 0.004.
Dielectric loss tan δ seems to be a little larger than the elastic loss tan φ.
e. Mechanically converted and stored energy = 100 W × k332 = 49 W
Among the stored mechanical energy, 1/3 ∼ 1/4 is usually spent out. Maximum output
mechanical energy is 1/2 of the stored energy = 25 W.
When this procedure is made quasistatically (off-resonance), the loss is the dielectric loss
portion (0.4%) = 0.4 W. When this procedure is made at this PZT rod’s resonance, the heat is
originated from the elastic loss, because the applied field is small, but the strain/stress excited in
the sample is large. This situation is illustrated in Figure 7.5.
7.1.2.3 Relaxor Ferroelectrics

Relaxor ferroelectrics can be prepared either in polycrystalline form or as single crystals. They dif-
fer from the previously mentioned normal ferroelectrics in that they exhibit a broad phase transition
from the paraelectric to ferroelectric state, a strong frequency dependence of the dielectric constant
(i.e., dielectric relaxation) and a weak remanent polarization. Lead-based relaxor materials have
complex disordered perovskite structures.
Relaxor-type electrostrictive materials, such as those from the lead magnesium niobate–lead
titanate, Pb(Mg1/3Nb2/3)O3–PbTiO3 (or PMN–PT) solid solution are highly suitable for actuator
applications. This relaxor ferroelectric also exhibits an induced piezoelectric effect. That is, the
electromechanical coupling factor kt varies with the applied DC bias field. As the DC bias field
increases, the coupling increases and saturates. Since this behavior is reproducible, these materials
can be applied as ultrasonic transducers which are tunable by the bias field.12
Recently, single-crystal relaxor ferroelectrics with the MPB composition have been developed
which show tremendous promise as ultrasonic transducers and electromechanical actuators. Single
crystals of Pb(Mg1/3Nb2/3)O3 (PMN), Pb(Zn1/3Nb2/3)O3 (PZN) and binary systems of these materials
combined with PbTiO3 (PMN–PT and PZN–PT) exhibit extremely large electromechanical cou-
pling factors.13,14 Large coupling coefficients and large piezoelectric constants have been found for
crystals from the MPBs of these solid solutions. PZN–8%PT single crystals were found to possess
a k33 value as high as 0.94 for the (001) crystal cuts; this is very high compared to the k33 of conven-
tional PZT ceramics of around 0.70 ∼ 0.80.
7.1.2.4 Polymers
PVDF or PVF2 is piezoelectric when stretched during fabrication. Thin sheets of the cast polymer
are then drawn and stretched in the plane of the sheet, in at least one direction, and frequently

[CH2CF2]n
Carbon
Fluorine
Z
Hydrogen
Y
X
FIGURE 7.6 Structure of PVDF.
also in the perpendicular direction, to transform the material to its microscopically polar phase.
Crystallization from the melt forms the nonpolar α-phase, which can be converted into the polar
β-phase by a uniaxial or biaxial drawing operation; the resulting dipoles are then reoriented through
electric poling (see Figure 7.6).
Large sheets can be manufactured and afterward thermally formed into complex shapes. The
copolymerization of vinilydene difluoride with trifluoroethylene (TrFE) results in a random copo-
lymer (PVDF-TrFE) with a stable, polar β-phase. This polymer does not need to be stretched; it
can be poled directly as formed. A thickness-mode coupling coefficient of 0.30 has been reported.
Piezoelectric polymers have the following characteristics: (a) small piezoelectric d constants (for
actuators) and large g constants (for sensors), due to small permittivity, (b) lightweight and soft
elasticity, leading to good acoustic impedance matching with water or the human body, (c) a low
mechanical quality factor Q M, allowing for a broad resonance band width. Such piezoelectric poly-
mers are used for directional microphones and ultrasonic hydrophones.
7.1.2.5 Composites
Piezocomposites comprised of a piezoelectric ceramic and a polymer phase are promising materials
because of their excellent and readily tailored properties. The geometry for two-phase composites
can be classified according to the dimensional connectivity of each phase into 10 structures; 0-0,
0-1, 0-2, 0-3, 1-1, 1-2, 1-3, 2-2, 2-3, and 3-3.15 A 1-3 piezocomposite, such as the PZT-rod/polymer
composite is the most promising candidate. The advantages of this composite are high coupling
factors, low acoustic impedance, good matching to water or human tissue, mechanical flexibility,
broad bandwidth in combination with a low mechanical quality factor, and the possibility of mak-
ing undiced arrays by structuring the electrodes. The thickness-mode electromechanical coupling
of the composite can exceed the kt (0.40 ∼ 0.50) of the constituent ceramic, approaching almost the
value of the rod-mode electromechanical coupling, k33 (0.70 ∼ 0.80) of that ceramic.16 The electro-
mechanical coupling factor of the composites is much better than polymer piezoelectrics. Acoustic
impedance is the square root of the product of its density and elastic stiffness. The acoustic match to
tissue or water (1.5 Mrayl) of the typical piezoceramics (20 ∼ 30 Mrayl) is significantly improved by
forming a composite structure, that is, by replacing some of the heavy, stiff ceramic with a light, soft
polymer. Piezoelectric composite materials are especially useful for underwater sonar and medical
diagnostic ultrasonic transducer applications. The details of these applications are described in
Chapter 10.
7.1.2.6 Thin Films

Both zinc oxide (ZnO) and aluminum nitride (AlN) are simple binary compounds with a Wurtzite-
type structure, which can be sputter-deposited as a c-axis oriented thin film on a variety of sub-
strates. ZnO has large piezoelectric coupling and thin films of this material are widely used in
bulk acoustic and SAW devices. The fabrication of highly oriented (along c) ZnO films have been
studied and developed extensively. However, the performance of ZnO devices is limited, due to their

low piezoelectric coupling (20% ∼ 30%). PZT thin films are expected to exhibit higher piezoelectric
properties. At present the growth of PZT thin films is being carried out for use in microtransducers
and microactuators. Refer to Section 3.3.6 for the thin film preparation techniques.
7.2 PRESSURE SENSORS/ACCELEROMETERS/GYROSCOPES

One of the very basic applications of piezoelectric ceramics is a gas igniter. The very high voltage
generated in a piezoelectric ceramic under applied mechanical stress can cause sparking and ignite
the gas (Figure 7.7). There are two means to apply the mechanical force, either by a rapid, pulsed
application or by a more gradual, continuous increase.
EXAMPLE PROBLEM 7.4

From the experimental data shown in Figure 7.7b, can you estimate the length L of the piezocer-
amic rod in Figure 7.7a?
Solution
If you know the relationship between the length L and the mechanical resonance frequency fr:
fr μ 1/L, and that 10 mm roughly corresponds to 100 kHz in PZT, you can estimate the rod length.
From the output voltage ringing, the resonance period is roughly estimated to be 30 μs, or a reso-
nance frequency of about 33 kHz, leading to a length L = 30 mm.
Piezoelectric ceramics can be employed as stress sensors and acceleration sensors, because of
the direct piezoelectric effect. Figure 7.8 shows a three-dimensional (3-D) stress sensor designed
by Kistler. By combining an appropriate number of quartz crystal plates (extensional and shear
types), the multilayer (ML) device can detect 3-D stresses.17
15
Output voltage (kV)
A 10
Electrode
L 5
Polarization
0 20 40 60
(a) (b) Time (μs)
FIGURE 7.7 (a) Gas igniter and (b) output voltage.
X
± ± ± ± ± 2
Z Y ± ± ± ± X
± ± ± ± ± 3
± ± ± ± Z
± ± ± ± ± 4
Y
1 ± ± ± ±
FIGURE 7.8 3-D stress sensor (by Kistler). (From Kistler, Stress Sensor, Production Catalog, Switzerland).

EXAMPLE PROBLEM 7.5

An accelerometer was fabricated using a piezoelectric ceramic disk (piezoelectric voltage con-
stant: g; thickness: h; and area: S) and a block mass (mass: M; and area: S) bonded together. See
Figure 7.9. This accelerometer was installed on a base excited by D 0 sin ωt (D 0: the maximum
displacement). Calculate the peak output voltage of the piezosensor.
Solution
Displacement D = D 0 sin ωt provides the acceleration:
d 2 D /dt 2 = ω 2 D0 sin ωt. (P7.5.1)
The stress applied to the piezodisk is given by
X = M (d 2 D /dt 2 )/S = − (ω 2D0 M /S ) sin ωt, (P7.5.2)
and the electric field generated is
E = gX = − (ω 2D0 Mg / S )sin ωt. (P7.5.3)
Thus, the peak output voltage is provided by
V0 = hEmax = ω 2 D0 Mgh/S. (P7.5.4)
Figure 7.10 shows a cylindrical gyroscope commercialized by NEC-Tokin (Japan).18 The cylinder
has six divided electrodes, one pair of which are used to excite the fundamental bending vibration
mode, while the other two pairs are used to detect the acceleration. When the rotational acceleration
is applied about the axis of this gyro, the voltage generated on the electrodes is modulated by the
Coriolis force. By subtracting the signals between the two sensor electrode pairs, a voltage directly
proportional to the acceleration can be obtained.
Mass M
Piezoelectric disk
D0 sin ωt Base
FIGURE 7.9 Basic structure of an accelerometer.
Vibrator
Lead Holder
Support
FIGURE 7.10 Cylindrical gyroscope (by NEC-Tokin).

Stress sensing rod
25 mm
PMN Bimorph
Brass box
C2 C1
Acrylic base
R1
r1
–
Oscillator +
1 kHz
–
R2 Peak/voltmeter
+
r2
–
+
Differential amplifier
FIGURE 7.11 Electrostrictive bimorph-type stress sensor.
The converse electrostrictive effect—the stress dependence of the permittivity—is also used in
stress sensors.19 A bimorph structure provides superior stress sensitivity and temperature stabil-
ity. Figure 7.11 shows one such measuring system, which subtracts the static capacitances of two
dielectric ceramic plates. The capacitance changes of the top and bottom plates have opposite signs
for uniaxial (up-down) stress and the same sign for temperature deviation. The response speed
is limited by the capacitance measuring frequency to about 1 kHz. Unlike piezoelectric sensors
(which is superior for high-frequency measurement), electrostrictive sensors are effective in the
low-frequency range, especially near DC.
Lab Demonstration 7.1: Direct Piezoelectricity (Figure 7.12)
1. Prepare a piezoelectric component such as a PZT ceramic disk, a ML, etc. If you do not
have them, you may find a bimorph speaker in your old electronic equipment (a mobile
phone or a laptop computer). Then, prepare a voltage/current meter.
2. Connect a positive (red color) and a negative lead wire (white/black color) of the piezo-
component to a positive/negative connector of the voltage meter. The lead wire color
corresponds to the original poling direction of the PZT ceramic.
3. Using your fingers, increase/decrease the compression on the piezocomponent, and
demonstrate positive/negative voltage/current, respectively.
4. Using an unidentified (no information on the polarization direction) piezocomponent, ask
the students to identify the polarization direction.

Multilayer
piezocomponent
FIGURE 7.12 Direct piezoelectric effect demonstration with a piezocomponent and a voltage/current meter.
7.3 PIEZOELECTRIC VIBRATORS/ULTRASONIC TRANSDUCERS

7.3.1 PIEZOELECTRIC RESONANCE
7.3.1.1 Piezoelectric Equations
When an electric field is applied to a piezoelectric material, deformation (ΔL) or strain (ΔL/L)
arises. When the field is alternating, mechanical vibration is generated, and if the drive frequency
is adjusted to a mechanical resonance frequency of the device, large resonating strain is generated.
This phenomenon can be explained by the strain amplification due to accumulating input energy
with time (amplification in terms of time), and is called piezoelectric resonance. The amplification
factor is proportional to the mechanical quality factor Q M. Piezoelectric resonance is very useful for
realizing energy trap devices, actuators, etc. The theoretical treatment is as follows.
If the applied electric field and the generated stress are not large, the stress X and the dielectric
displacement D can be represented by the following equations:
xi = sijE X j + dmi Em , i, j = 1,2,…,6; m, k = 1,2,3 (7.20)
Dm = dmi Xi + ε Ek .
X
mk (7.21)
These are called the piezoelectric equations. The number of independent parameters for the low-
est symmetry trigonal crystal is 21 for sijE, 18 for dmi, and 6 for εmk X . The number of independent
parameters decreases with increasing crystallographic symmetry. Concerning the polycrystalline

ceramics, the poled axis is usually denoted as the z-axis and the ceramic is isotropic with respect
to this z-axis (Curie group C∞v (∞m)). The number of nonzero matrix elements in this case is 10
E , s E , s E , s E , s E , d , d , d , ε X, and ε X). Refer to Section 2.1.
(s11 12 13 33 44 31 33 15 11 33
7.3.1.2 Electromechanical Coupling Factor

Next, let us introduce the electromechanical coupling factor k, which corresponds to the rate of
electromechanical transduction. The internal energy U of a piezoelectric vibrator is given by sum-
mation of the mechanical energy UM(= ∫x dX) and the electrical energy UE(= ∫D dE). U is calculated
as follows, when linear relations (Equations 7.20 and 7.21) are applicable:
U = UM + UE
⎡ ⎤ ⎡ ⎤
= ⎢(1 2 )
⎢
⎣
∑ i ,j
sijE X j Xi + (1 2 ) ∑ m ,i
dmi Em Xi ⎥ + ⎢ (1 2 )
⎥
⎦ ⎢⎣
∑ m ,i
∑
dmi Xi Em + (1 2 )
k ,m
ε mk
X
Ek Em ⎥
⎥⎦
= U MM + 2U ME + U EE
= (1 2 )∑s i ,j
E
ij X j Xi + 2 ⋅ (1 2 )∑d m ,i
mi Em Xi + (1 2 ) ∑εk ,m
X
mk Ek Em . (7.22)

The s and ε terms represent purely mechanical and electrical energies (UMM and UEE), respectively,
and the d term denotes the energy transduced from electrical to mechanical energy or vice versa
through the piezoelectric effect (UME). The electromechanical coupling factor k is defined by
k = U ME U MM ⋅ U EE . (7.23)
Note that this definition is equivalent to the definition provided in Section 7.1.1:
k 2 = U ME /U E = (Stored mechanical energy/Input electrical energy)
or
k 2 = U ME /U E = (Stored electrical energy/Input mechanical energy).
The k value varies with the vibration mode (even in the same ceramic sample), and can have a posi-
tive or negative value (see Table 7.3). Referring to Table 7.2, k31/k33 ratio around 0.47 originates from
the d31/d33 ratio around 0.43. The k value is primarily governed by the contributing piezoelectric d
constant for that vibration mode.
EXAMPLE PROBLEM 7.6

Calculate the electromechanical coupling factor kij of a piezoelectric ceramic vibrator for the
following vibration modes (see Figure 7.13):
a. Longitudinal length extension mode (//E): k33

b. Planar extension mode of the circular plate: kp
Solution
Using the expression
U = U MM + 2U ME + U EE
∑s
= (1 2 )
i ,j
E
ij ∑d
X j Xi + 2 ⋅(1 2 )
m ,i
mi ∑ε
Em Xi + (1 2 )
k ,m
X
mk Ek Em , (P7.6.1)
the electromechanical coupling factor is given by
k = U ME U MM ⋅ U EE . (P7.6.2)
X
Electrode
Polarization
direction
(a) (b)
FIGURE 7.13 (a) Longitudinal length extension and (b) planar extensional vibration modes of piezoelectric
vibrators.

TABLE 7.3
Several Shapes of the Piezoelectric Resonator and their EM Coupling Factors
Coupling Elastic Boundary
Factor Conditions Resonator Shape Definition
A k31 X1 ≠ 0, X2 = X3 = 0 3 d31
x1 ≠ 0, x2 ≠ 0, x3 ≠ 0 1
s11E ε 33
X
B k33 X1 = X2 = 0, X3 ≠ 0 3 d33
x1 = x2 ≠ 0, x3 ≠ 0
s 33E ε 33
X
Fundamental Mode
C kp X1 = X2 ≠ 0, X3 = 0
2
x1 = x2 ≠ 0, x3 ≠ 0 k31
1− σ
3
Fundamental Mode
D kt X1 = X2 ≠ 0, X3 ≠ 0
ε 33
x
x1 = x2 = 0, x3 ≠ 0 k33 D
c33
3
Thickness Mode
mode
E k′p X1 = X2 ≠ 0, X3 ≠ 0 3 kp − Ak33
x1 = x2 ≠ 0, x3 = 0
1 − A2 1 − k33
2
Radial Mode
F k′31 X1 ≠ 0, X2 ≠ 0, X3 = 0 3 2 k31 1+ σ
x1 ≠ 0, x2 = 0, x3 ≠ 0 1
1 − k31
2 1− σ
Width Mode
G k″31 X1 ≠ 0, X2 = 0, X3 ≠ 0 3 k31 − Bk33
x1 ≠ 0, x2 ≠ 0, x3 = 0 1
2 1 − k33
2
Width Mode
H k33″′ X1 ≠ 0, X2 ≠ 0, X3 ≠ 0 1
2 (kp − Ak33 )2
x1 ≠ 0, x2 = 0, x3 = 0 − (k31 − Bk33 )2
1 − A2
3
Thickness Mode 1 − k33
2
− (k31 − Bk33 )2
I k′33 X1 ≠ 0, X2 = 0, X3 ≠ 0 1 k33 − Bk31

x1 = 0, x2 ≠ 0, x3 ≠ 0 3
(1 − B2 )(1 − k31
2
)
Width Mode
J k24 = k15 X1 = X2 = X3 = 0, X4 ≠ 0 d15
x1 = x2 = x3 = 0, x4 ≠ 0 ε11
X E
s44
E E
2 s13 s13
Here: A = , B=
E
s33 E
(s11 + s12
E
) E E
s11 s33

a. The relating equations for this k33 mode are
x3 = s33
E
X3 + d33 E3
D3 = d33 X3 + ε 33
X
E3
→UME comes from the d33 term as (1/2)d33E3X3, UMM comes from the sE33term as (1/2)sE33 X23,
and UEE comes from the ε33
X term as (1/2) ε X E2. Thus
33 3
⎛ 1⎞ E 2 ⎛ 1⎞ X 2
k33 = (1 2 )d33 E3 X3 ⎜⎝ 2 ⎟⎠ s33 X3 ⋅ ⎜⎝ 2 ⎟⎠ ε 33 E3
= d33 E
s33 ⋅ ε 33
X
. (P7.6.3)
b. The relating equations for the kp mode are following three, including 2-D x1 and x2
equations:
x1 = s11E X1 + s12E X 2 + d31E3
x2 = s12E X1 + s22
E
X 2 + d32 E3
D3 = d31 X1 + d32 X 2 + ε 33
X
E3 .
Assuming axial symmetry, s11 E = s E , d = d , and X = X (=X ), the above equations are
22 31 32 1 2 p
transformed to the following two equations:
x1 + x2 = 2(s11E + s12E ) X p + 2d31E3
D3 = 2d31 X p + ε 33
X
E3 .
→UME comes from the d31 term as (1/2) . 2d31E3Xp, UMM comes from the 2(s11 E + s E ) term as
12
(1/2) . 2(s11
E + s E ) X2 , and U
12 p EE comes from the εX term as (1/2) εX E2. Thus
33 33 3
k33 = 2d31 (E
2 s11 + s12
E
)
⋅ ε 33
X
= ⎡⎢ d31 X ⎤ 2 2
E
s11 ⋅ ε 33 ⋅ = k31 ⋅ , (P7.6.4)
⎣ ⎦⎥ 1 − σ 1− σ
where σ is Poisson’s ratio given by
σ = − s12
E E
/ s11 . (P7.6.5)
7.3.1.3 Longitudinal Vibration Mode

Let us consider the longitudinal mechanical vibration of a piezoceramic (polycrystalline, not a single
crystal) plate through the transverse piezoelectric effect (d31) as shown in Figure 7.14. If the polarization
is in the z-direction and x–y planes are the planes of the electrodes, the extensional vibration in the x
direction is represented by the following dynamic equation (when the length L is more than 4 ~ 6 times of
the width w or the thickness b, we can neglect the coupling modes with width or thickness vibrations):
(∂ 2u/∂t 2 ) = F = (∂X11 /∂x ) + (∂X12 /∂y) + (∂X13 /∂z ), (7.24)

z y
b Pz x
0
L
FIGURE 7.14 Longitudinal vibration through the transverse piezoelectric effect (d31) in a rectangular plate
(L >> w >> b).
where u is the displacement of the small volume element in the ceramic plate in the x-direction. The
relations between stress, electric field (only Ez exists), and the induced strain are given by
x1 = s11
E
X1 + s12
E
X 2 + s13
E
X3 + d31E3 ,
x2 = s12
E
X1 + s11
E
X 2 + s13
E
X3 + d31E3 ,
x3 = s13
E
X1 + s13
E
X 2 + s33
E
X3 + d33 E3 ,
x4 = s44
E
X4 ,
x5 = s44
E
X5,
x6 = 2(s11
E
− s12
E
) X6 . (7.25)
When the plate is very long and thin, X2 and X3 may be set equal to zero through the plate. Since
shear stress will not be generated by the electric field Ez (=E3), Equation 7.25 is reduced to only one
equation:
X1 = x1 /s11
E
− (d31 /s11
E
)Ez . (7.26)
Introducing Equation 7.26 into Equation 7.24, and allowing for x1 = ∂u/∂x (nonsuffix x corresponds
to the Cartesian coordinate, and x1 is the strain along the 1 (x) direction) and ∂Ez /∂x = 0 (due to the
equal potential on each electrode), leads to a harmonic vibration equation:
−ω 2ρs11
E
u = ∂ 2u/∂x 2, (7.27)
where
ω is the angular frequency of the drive field
ρ is the density
Substituting a general solution u = u1 (x)ejω t + u2 (x)e−jω t into Equation 7.26, and with the boundary
condition X1 = 0 at x = 0 and L (sample length) (due to the mechanically free condition at the plate
end), the following solution can be obtained:
∂u ∂x = x1 = d31Ez ⎡⎣sin ω (L − x ) v + sin (ωx v )⎤⎦ sin (ωL v )
⎛ ⎡ ω (L − 2 x )⎤ ⎞
⎜ cos ⎢ ⎥⎟
⎣ 2v ⎦⎟ ,
= d31Ez ⎜ (7.28)
⎜ ⎛ ωL ⎞ ⎟
⎜ cos ⎜ ⎟ ⎟
⎝ ⎝ 2v ⎠ ⎠

where, v is the sound velocity in the piezoceramic along x direction, which is given by
v =1 ρs11
E
. (7.29)
When the specimen is utilized as an electrical component such as a filter or a vibrator, the electrical
impedance [(applied voltage/induced current) ratio] plays an important role. The current flow into
the specimen is described by the surface charge increment, ∂D3/∂t, and the total current is given by
L L
∫
i = jωw D3 dx = jωw
0
∫ (d
0
31 X1 + ε 33
X
)
Ez dx
⎣ ∫ {
= jωw ⎡ d31 x1 s11
0
E
− d31 s11
E
Ez( ) }+ ε X
33 Ez ⎤ dx.
⎦
(7.30)
Using Equation 7.28, the admittance for the mechanically free sample is calculated to be
(1 Z ) = (i v ) = (i Ezt )
⎡ ⎛ ⎡ ω (L − 2 x )⎤ ⎞ ⎤
L⎢ ⎜ cos ⎢ ⎥⎟ ⎥
⎦ ⎟ E + ⎡ ε X − ⎛ d31 ⎞ ⎤ E ⎥ dx,
∫( )
⎢ 2 E ⎣ 2v 2
= ( j ωw Ez t ) ⎢ d31 s11 ⎜ ⎢ E ⎟⎥ z⎥
⎜ ⎛ ωL ⎞ ⎟ z ⎣⎢ 33 ⎜⎝ s11 ⎠ ⎦⎥ ⎥
0 ⎢ ⎜ cos ⎜⎝ 2v ⎟⎠ ⎟
⎢ ⎝ ⎠ ⎥
⎣ ⎦
= ( jωwL t )ε 33
LC ⎡
⎣
1 + d31
2
(
ε 33
LC E
)
s11 tan (ωL 2v) (ωL 2v)⎤ ,
⎦
(7.31)
where
w is the width
L is the length
t is the thickness of the rectangular piezosample
V is the applied voltage
εLC
33 is the permittivity in a longitudinally clamped sample, which is given by
ε 33
LC
= ε 33
X 2
(
− d31 E
s11 . ) (7.32)
The piezoelectric resonance is achieved where the admittance becomes infinite or the impedance is
zero. The resonance frequency f R is calculated from Equation 7.31 (by putting ωL/2v = π/2), and the
fundamental frequency is given by
fR = ω R 2π = v 2 L = 1 2 L ρs11
E
(
. ) (7.33)
On the other hand, the antiresonance state is generated for zero admittance or infinite impedance:
(ω A L /2v) cot (ω A L /2v) = − d31

2
/ε 33 s11 = − k31
LC E 2
/(1 − k31
2
). (7.34)
The final transformation is provided by the definition
k31 = d31 E
s11 ⋅ ε 33
X
. (7.35)

Resonance Antiresonance
Low coupling High coupling
m=1 m=1
m=2 m=2
FIGURE 7.15 Strain distribution in the resonant or antiresonant state.
The resonance and antiresonance states are both mechanical resonance states with amplified strain/
displacement states, but they are very different from the electrical/driving viewpoints. The mode
difference is described by the following intuitive model. In a high electromechanical coupling mate-
rial with k almost equal to 1, the resonance or antiresonance states appear for tan(ωL/2v) = ∞ or 0
[i.e., ωL/2v = (m − 1/2)π or mπ (m: integer)], respectively. The strain amplitude x1 distribution for
each state (calculated using Equation 7.28) is illustrated in Figure 7.15. In the resonance state, large
strain and capacitance changes (called motional capacitance) are induced, and the current can
easily flow into the device. To the contrary, at the antiresonance, the strain induced in the device
compensates completely, resulting in no capacitance change, and the current cannot flow easily into
the sample. Thus, for a high-k material the first antiresonance frequency fA should be twice as large
as the first resonance frequency f R.
It is notable that both resonance and antiresonance states are in the mechanical resonance, which
can create large strain in the sample under minimum input electrical energy. The stress X1 at the
plate ends (x = 0 and L) is supposed to be zero in both cases. However, though the strain x1 at the
plate ends is zero for the resonance, the strain x1 is not zero for the antiresonance. This means that
there is only one vibration node at the plate center for the resonance (top-left in Figure 7.15), and
there are additional two nodes at both plate ends for the antiresonance (top-right in Figure 7.15). The
reason is from the antiresonance drive, i.e., high voltage/low current drive due to the high imped-
ance. The converse piezoeffect strain under E via d31 (uniform strain in the sample) superposes
on the mechanical resonance strain distribution (distributed strain with nodes in the sample), two
strains of which have exactly the same level theoretically at the antiresonance for k31 = 1.
In a typical case, where k31 = 0.3, the antiresonance state varies from the previously mentioned
mode and becomes closer to the resonance mode (top-center in Figure 7.15). The low-coupling mate-
rial exhibits an antiresonance mode where the capacitance change due to the size change (motional
capacitance) is compensated partially by the current required to charge up the static capacitance
(called damped capacitance). Thus, the antiresonance frequency fA will approach the resonance
frequency f R. You may also consider that the two additional nodes at the plate ends (top right in
Figure 7.15) are now outside of the plates.
The general procedure for calculating the electromechanical parameters (k31, d31, s11 E , and ε X )
33
from the impedance spectrum measurement is described below:
1. The sound velocity v in the specimen is obtained from the resonance frequency f R (refer to
Figure 7.16), using Equation 7.33: f R = v/2L.
2. Knowing the density ρ, the elastic compliance s11E can be calculated from the sound veloc-
ity v: v = 1 ρs11 .
E
3. The electromechanical coupling factor k31 is calculated from the v value and the antireso-
nance frequency fA through Equation 7.34. Especially in low-coupling piezoelectric mate-
rials, the following approximate equation is useful:
2
k31 /(1 − k31
2
) = (π 2 /4)( Δf / fR ) ( Δf = fA − fR ). (7.36)

k33 = 0.70 fA = 465 kHz

= 1.3 fR
Impedance
fR = 360 kHz
(a) Frequency
k33 = 0.90
fA = 584 kHz
Impedance
= 2 fR
1/ωC0
fR = 295 kHz
(b) Frequency
FIGURE 7.16 Impedance curves for (a) a reasonable k material (PZT 5 H, k33 = 0.70), and (b) a high-k
material (PZN–PT single crystal, k33 = 0.90).
4. Knowing the permittivity ε33 X from the independent measurement under an off-resonance
condition, the d31 is calculated from k31 through Equation 7.35:
k31 = d31 E
s11 ⋅ ε 33
X
.
Figure 7.16a and b compares observed impedance curves of rod-shaped samples for a typical k
material (PZT 5 H, k33 = 0.70) and a high-k material (PZN–PT single crystal, k33 = 0.90). Note a
large separation between the resonance and antiresonance peaks in the high-k material, leading to
the condition almost fA = 2f R. To the contrary, a regular PZT sample exhibits fA = 1.3f R. The band-
width of the piezo-transducer is defined by (Δ f /f R), leading to 100% for the PZN–PT rod sample
and only 30% for PZT 5 H.
EXAMPLE PROBLEM 7.7

Knowing the above experimental result: fA = 1.3f R, calculate the k33 of this PZT 5 H.
Solution
2
k33 /(1 − k33
2
) = (π 2 /4) ( Δf / fR ) = (π 2 /4) × 0.3 = 0.74,
thus, k33 = 0.65.
7.3.2 EQUIVALENT CIRCUITS OF PIEZOELECTRIC VIBRATORS

The dynamic equation for a mechanical system composed of a mass, a spring, and a damper illus-
trated in Table 7.4 Left is expressed by
M (d 2u/dt 2 ) + ζ(du/dt ) + cu = F (t ), (7.37)
where
u is the displacement of a mass M
c is a spring constant

Table 7.4
Equivalency between a Mechanical and an Electrical System, Composed
of M, c, z; L, C, and R
Mechanical Electrical
Displacement u Charge Q
Force F Voltage V L
c Mass M Inductance L
Damping ζ Resistance R
Spring constant c Inverse Capacitance 1/C C
ζ
R
M
ζ is a damping constant of a dashpot

F is the external force
On the other hand, the dynamic equation for an electrical circuit composed of an inductance L,
a capacitance C, and a resistance R illustrated in Table 7.4 (right) is expressed by
L (d 2Q /dt 2 ) + R(dQ /dt ) + (1/C )Q = V (t ), (7.38)
where
Q is the charge
V is the voltage
Using the duality as summarized in Table 7.4, equivalent circuits of piezoelectric vibrators can be
generated easily by mechanical and electrical engineers through mechanical or electrical point of
view.
Now, the equivalent circuit for the piezoelectric transducer is represented by a combination of
L, C, and R. Figure 7.17a shows an equivalent circuit for the resonance state, which has very low
impedance. Cd (damped capacitance) corresponds to the electrostatic capacitance, and the compo-
nents LA and CA in a series resonance circuit are related to the piezoelectric motion. The subscript A
and the B below denote “A-type” and “B-type” resonance modes, which mean equally “resonance”
Gd Cf
LA
Cd CA
RA GB LB
CB
(a) (b)
FIGURE 7.17 Equivalent circuit of a piezoelectric device for (a) the resonance and (b) the antiresonance
states.

and “antiresonance” modes. For example, in the case of the longitudinal vibration of the above
rectangular plate through d31, these components are represented by
LA = (ρ/8)( Lb/w)(s11
E2 2
/d31 ), (7.39)
CA = (8/π2 )( Lw/b)(d31
2 E
/s11). (7.40)
Note that the inductance LA is directly proportional to the density ρ, the inverse capacitance CA is
directly proportional to the elastic compliance s11E (i.e., spring constant), while the component RA
corresponds to the mechanical loss [more precisely, a combination of dielectric loss (imaginary
permittivity constant), elastic loss (imaginary elastic compliance), and piezoelectric loss (imaginary
d constant)].
In contrast, the equivalent circuit for the antiresonance state of the same actuator, which has high
impedance is shown in Figure 7.17b.
7.3.3 PIEZOELECTRIC VIBRATORS

In the use of mechanical vibration devices such as filters or oscillators, the size and shape of a
device are very important, and both the vibrational mode and the ceramic material must be con-
sidered. The resonance frequency of the bending mode in a centimeter-size sample ranges from
100 to 1000 Hz, which is much lower than that of the thickness mode (1 cm→100 kHz). For these
vibrator applications the piezoceramic should have a high mechanical quality factor (Q M) rather
than a large piezoelectric coefficient d; that is, hard piezoelectric ceramics are preferred.
For speakers or buzzers, audible by humans, devices with a rather low resonance frequency are
used (100 Hz−2 kHz range). Examples are a unimorph consisting of one piezoceramic plate bonded
with a metallic shim, a bimorph consisting of two piezoceramic plates bonded together, and a piezo-
electric fork consisting of a piezodevice and a metal fork. A piezoelectric buzzer design is shown in
Figure 7.18a, which has merits such as high efficiency, compact size, and long life.
Figure 7.18b shows a state-of-art speaker with only 0.7 mm ultrathin thickness and 0.4 g weight.20
The power consumption is only 1/5 ~ 2/3 compared to electromagnetic types. The piezospeaker has
wide frequency range and high sound pressure, and in particular, no adverse effect on credit cards,
which is important nowadays.
Elastic vibrator Piezoceramic
(a)
FIGURE 7.18 Basic structure of a piezoelectric buzzer (a) and a state-of-art ML ceramic speaker (b).
(From Watabe, Y., Bull. Ceram. Soc. Jpn., 42, 396, 2007. With permission.)

7.3.4 ULTRASONIC TRANSDUCERS

Ultrasonic waves are now used in various fields. The sound source is made from piezoelectric
ceramics, as well as magnetostrictive materials. Piezoceramics are generally superior in efficiency
and in size to magnetostrictive materials. In particular, hard piezoelectric materials with a high Q M
are preferred because of high-power sound generation without heat generation. A liquid medium
is usually used for sound energy transfer. Ultrasonic washers, ultrasonic microphones, and sonars
for short-distance remote control, underwater detection, finding fish, and nondestructive testers are
typical applications. Ultrasonic scanning detectors are useful in medical electronics for clinical
applications ranging from diagnosis to therapy and surgery.
7.3.4.1 Ultrasonic Imaging

One of the most important applications is based on ultrasonic echo field.21,22 Ultrasonic transduc-
ers convert electrical energy into mechanical form when generating an acoustic pulse and convert
mechanical energy into an electrical signal when detecting its echo. The transmitted waves propa-
gate into a body and echoes are generated which travel back to be received by the same transducer.
These echoes vary in intensity according to the type of tissue or body structure, thereby creating
images. An ultrasonic image represents the mechanical properties of the tissue, such as density and
elasticity. We can recognize anatomical structures in an ultrasonic image since the organ boundar-
ies and fluid-to-tissue interfaces are easily discerned. The ultrasonic imaging process can also be
done in real time. This means we can follow rapidly moving structures such as the heart without
motional distortion. Furthermore, ultrasound is one of the safest diagnostic imaging techniques. It
does not use ionizing radiation like x-rays and thus is routinely used for fetal and obstetrical imag-
ing. Useful areas for ultrasonic imaging include cardiac structures, the vascular systems, the fetus,
and abdominal organs such as liver and kidney. In brief, it is possible to see inside the human body
without breaking the skin with the help of a beam of ultrasound.
Figure 7.19 shows the basic ultrasonic transducer geometry. The transducer is mainly composed
of matching piezoelectric material and backing layers.23 One or more matching layers are used to
increase sound transmissions into tissues. The backing is added to the rear of the transducer in order
to damp the acoustic backwave and to reduce the pulse duration. Piezoelectric materials are used
to generate and detect ultrasound. In general, broadband transducers should be used for medical
ultrasonic imaging. The broad bandwidth response corresponds to a short pulse length, resulting
in better axial resolution. Three factors are important in designing broad bandwidth transducers;
acoustic impedance matching, a high electromechanical coupling coefficient of the transducer, and
electrical impedance matching. These pulse echo transducers operate based on thickness mode
resonance of the piezoelectric thin plate. Further, a low planar mode coupling coefficient, kp, is
beneficial for limiting energies being expended in nonproductive lateral mode. A large dielectric
constant is necessary to enable a good electrical impedance match to the system, especially with
tiny piezoelectric sizes.
Piezoelectric element
Backing Matching layer
Ultrasonic
beam
Input
pulse
FIGURE 7.19 Basic transducer geometry for acoustic imaging applications.

W
L
Polarization direction
T
(a) Vibrator element
Piezoelectric vibrator
Backing
(b) Structure of an array-type ultrasonic probe
FIGURE 7.20 Linear array-type ultrasonic probe.
There are various types of transducers used in ultrasonic imaging. The first type of transduc-
ers are mechanical sector transducers consist of single, relatively large resonators and can provide
images by mechanical scanning such as wobbling. The second type of transducers are multiple ele-
ment array transducers permit discrete elements to be individually accessed by the imaging system
and enable electronic focusing in the scanning plane to various adjustable penetration depths through
the use of phase delays. Two basic types of array transducers are linear and phased (or sector). A
linear array is a collection of elements arranged in one direction, producing a rectangular display
(see Figure 7.20). A curved linear (or convex) array is a modified linear array whose elements are
arranged along an arc to permit an enlarged trapezoidal field of view. The elements of these linear-
type array transducers are excited sequentially group by group with the sweep of the beam in one
direction. These linear array transducers are used for radiological and obstetrical examinations. On
the other hand, in a phased array transducer the acoustic beam is steered by signals that are applied
to the elements with delays, creating a sector display. This transducer is useful for cardiology appli-
cations where positioning between the ribs is necessary.
Figure 7.21 demonstrates the superiority of the PZN–PT single crystal to the PZT ceramic for the
medical imaging transducer application, developed by Toshiba Corporation.24 Conventionally, the
medical doctor is required to use two different frequency PZT probes, one at 2.5 MHz for checking
wider and deeper area, and a second one at 3.75 MHz for monitoring the specified area with a bet-
ter resolution. The PZN–PT single crystal (with very high k33 and kt) probe provides two additional
merits: (1) wide bandwidth—without changing the probe, the doctor can just switch the drive fre-
quency from 2.5 to 3.75 MHz, (2) strong signal—because of the high electromechanical coupling,
the receiving signal level is enhanced more than twice compared with the PZT probe.
7.3.4.2 Sonochemistry
Fundamental research on “Sonochemistry” is now rapidly ongoing. Using “cavitation” effect, toxic
materials such as dioxin and trichloroethylene can be easily transformed into innocuous materials at
room temperature. Ultrasonic distillation is also possible at room temperature for obtaining highly
concentrated Japanese “Sake.” Different from the regular boiling distillation, this new method
makes Sake at much higher alcoholic concentration while keeping gorgeous taste and fragrance.
Figure 7.22a shows the alcoholic concentration in the base solution and mist. This high-quality Sake
product is now commercially available.25

Comparison of the Doppler mode images
Higher resolution
(the wall of heart)
Greater penetration
(blood flow in the pulmonary vein)
3.75-MHz PZT 2.5-MHz PZT PZNT 91/9
ceramic probe ceramic probe single-crystal probe
PZNT probe’s performance includes both penetration of 2.5 MHz PZT
probe and resolution of 3.75 MHz PZT probe Toshiba
FIGURE 7.21 Ultrasonic imaging with the two PZT ceramic probes (left) and with the PZN–PT single
crystal probe (right). (Yamashita, Y., Private communication; Courtesy of Toshiba, 2004. With permission.)
High-power ultrasonics are also applicable to transdermal drug delivery. The Penn State
researchers are working with commercialization of this “needle-free” injection system of insulin by
using cymbal piezoactuators (see Figure 7.22b).26
7.3.5 RESONATORS/FILTERS
When a piezoelectric body vibrates at its resonant frequency it absorbs considerably more energy
than at other frequencies resulting in a dramatic decrease in the impedance. This phenomenon
enables piezoelectric materials to be used as a wave filter. A filter is required to pass a certain
selected frequency band or to block a given band. The bandwidth of a filter fabricated from a piezo-
electric material is determined by the square of the coupling coefficient k, that is, it is nearly pro-
2 /(1 − k 2 ) = (π2/4)(Δ f /f ), where Δf = f − f ,
portional to k2. The background is from the relation: k31 31 R A R
and the bandwidth is provided by Δf. Quartz crystals with a very low k value of about 0.1 can pass
very narrow frequency bands of approximately 1% of the center resonance frequency. On the other
hand, PZT ceramics with a planar coupling coefficient of about 0.5 can easily pass a band of 10%
of the center resonance frequency. The sharpness of the passband is dependent on the mechanical
quality factor Q M of the materials. Quartz also has a very high Q M of about 106 which results in a
sharp cutoff to the passband and a well-defined oscillation frequency.
A simple resonator is a thin piezo-disk, electroded on its plane faces and vibrating radially for
filter applications with a center frequency ranging from 200 kHz to 1 MHz and with a bandwidth of
several percent of the center frequency. For a frequency of 455 kHz the disk diameter needs to be
about 5.6 mm. However, if the required frequency is higher than 10 MHz, other modes of vibration
such as the thickness extensional mode are exploited, because of its smaller size. The trapped-energy-
type filters made from PZT ceramics have been widely used in the intermediate frequency range
for applications such as the 10.7 MHz FM radio receiver and transmitter. When the trapped-energy
phenomena are utilized, the overtone frequencies are suppressed. The plate is partly covered with

100
Ethanol mol concentration in mist (mol %)

Boiling distillation
50
10°C 30°C 50°C
0
0 50 100
Ethanol mol concentration in solution (mol %)
(a)
Illustration by Paul Perreault
(b)
FIGURE 7.22 (a) Room temperature distillation with high-power ultrasonic (From http://www.shumurie.
co.jp, Popular Mechanics, 180, 20, 2003. With permission.) and (b) transdermal insulin drug delivery system
using cymbal transducers. (Courtesy of Matsuura Brewer.)
electrodes of a specific area and thickness. The fundamental frequency of the thickness mode of
the ceramic beneath the electrode is less than that of the unelectroded portion, because of the extra
inertia of the electrode mass. The lower-frequency wave of the electroded region cannot propagate
into the unelectroded region. The higher-frequency overtones, however, can propagate away into
the unelectroded region. This is called the trapped-energy principle. Figure 7.23 shows a schematic
drawing of a trapped-energy filter. In this structure the top electrode is split so that coupling between

Electrode
Ceramic plate
Top Bottom
FIGURE 7.23 Schematic drawing of a trapped-energy filter.
the two parts will only be efficient at resonance. More stable filters suitable for telecommunication
systems have been made from single crystals such as quartz or LiTaO3.
7.4 SURFACE ACOUSTIC WAVE DEVICES

A SAW, also called a Rayleigh wave, is essentially a coupling between longitudinal and shear waves.
The energy carried by the SAW is confined near the surface. An associated electrostatic wave exists
for a SAW on a piezoelectric substrate, which allows electroacoustic coupling via a transducer. The
advantages of SAW technology are27,28:
1. The wave can be electroacoustically accessed and tapped at the substrate surface and its
velocity is approximately 104 times slower than an electromagnetic wave
2. The SAW wavelength is on the same order of magnitude as line dimensions produced by
photolithography and the lengths for both short and long delays are achievable on reason-
ably sized substrates
There is a very broad range of commercial system applications which include front-end and inter-
mediate frequency (IF) filters, community antenna television (CATV) and video cassette recorder
(VCR) components, synthesizers, analyzers, and navigators. In SAW transducers, finger (interdigi-
tal) electrodes provide the ability to sample or tap the wave and the electrode gap gives the relative
delay. A SAW filter is composed of a minimum of two transducers. A schematic of a simple SAW
bidirectional filter is shown in Figure 7.24. A bidirectional transducer radiates energy equally from
each side of the transducer. Energy which is not associated with the received signal is absorbed to
eliminate spurious reflection.
Various materials are currently being used for SAW devices. The most popular single-crystal
SAW materials are lithium niobate and lithium tantalate. The materials have different properties
depending on the cut of the material and the direction of propagation. The fundamental parameters
considered when choosing a material for a given device applications are SAW velocity, temperature
coefficients of delay (TCD), electromechanical coupling factor, and propagation loss. SAWs can be
SAW
Input Output
Interdigital electrode
Piezoelectric substrate
FIGURE 7.24 Fundamental structure of a SAW device.

generated and detected by spatially periodic, interdigital electrodes on the plane surface of a piezo-
electric plate. A periodic electric field is produced when an RF source is connected to the electrode,
thus permitting piezoelectric coupling to a traveling surface wave. If an RF source with a frequency,
f, is applied to the electrode having periodicity, d, energy conversion from an electrical to mechani-
cal form will be maximum when
f = f0 = vs /d, (7.41)
where
vs is the SAW velocity
f0 is the center frequency of the device
The SAW velocity is an important parameter determining the center frequency. Another important
parameter for many applications is temperature sensitivity. For example, the temperature stability
of the center frequency of SAW band-pass filters is a direct function of the temperature coefficient
for the velocity and the delay for the material used. The first-order temperature coefficient of delay
is given by
(1/τ) ⋅ (dτ /dT) = (1/L) ⋅ (dL /dT) − (1/vs ) ⋅ (dvs /dT ), (7.42)
where
τ = L/vs is the delay time
L is the SAW propagation length
The surface wave coupling factor, k2s , is defined in terms of the change in SAW velocity which
occurs when the wave passes across a surface coated with a thin massless conductor, so that the
piezoelectric field associated with the wave is effectively short-circuited. The coupling factor, ks2,
is expressed by
ks2 = 2(vf − vm )/vf , (7.43)
where
vf is the free surface wave velocity
vm is the velocity on the metallized surface
In actual SAW applications, the value of k2s relates to the maximum bandwidth obtainable and the
amount of signal loss between input and output, which determines the fractional bandwidth as a
function of minimum insertion loss for a given material and filter. Propagation loss is one of the
major factors that determines the insertion loss of a device and is caused by wave scattering at
crystalline defects and surface irregularities. Materials which show high electromechanical cou-
pling factors combined with small temperature coefficients of delay are generally preferred.
The free surface velocity, vf, of the material is a function of cut angle and propagation direction.
The TCD is an indication of the frequency shift expected for a transducer due to a temperature
change and is also a function of cut angle and propagation direction. The substrate is chosen based
on the device design specifications which include operating temperature, fractional bandwidth, and
insertion loss.
Piezoelectric single crystals such as 128°Y-X (128°-rotated-Y-cut and X-propagation)—LiNbO3
and X-112°Y (X-cut and 112°-rotated-Y-propagation)—LiTaO3 have been extensively employed as
SAW substrates for applications in video intermediate frequency (VIF) filters. A c-axis-oriented ZnO
thin film deposited on a fused quartz, glass, or sapphire substrate has also been commercialized for
SAW devices. Table 7.5 summarizes some important material parameters for these SAW materials.

TABLE 7.5
SAW Material Properties
Cut-Propagation TCD
Material Direction k2 (%) (ppm/C) V0 (m/s) er
Single crystal Quartz ST–X 0.16 0 3158 4.5
LiNbO3 128°Y–X 5.5 −74 3960 35
LiTaO3 X112°–Y 0.75 −18 3290 42
Li2B4O7 (110)–<001> 0.8 0 3467 9.5
Ceramic PZT-In(Li3/5W2/5)O3 1.0 10 2270 690
(Pb,Nd) (Ti,Mn,In)O3 2.6 <1 2554 225
Thin film ZnO/glass 0.64 −15 3150 8.5
ZnO/sapphire 1.0 −30 5000 8.5
A delay line can be formed from a slice of glass such as PbO- or K2O-doped SiO2 glass in
which the velocity of sound is nearly independent of temperature. PZT ceramic transducers are
soldered on two metallized edges of the slice of glass. The input transducer converts the electrical
signal to a shear acoustic wave which travels through the slice. At the output transducer the wave is
reconverted into an electrical signal delayed by the length of time taken to travel around the slice.
Such delay lines are used in color TV sets to introduce a delay of approximately 64 μs and are also
employed in videotape recorders.
Recent development on SAW actuator applications is interesting. A liquid transportation system
was developed by using a standing wave–type SAW device.29 A liquid droplet can be transported
by controlling the SAW wave.
7.5 MICROMASS SENSORS

7.5.1 BIOSENSOR
Quartz is not only for timers of clocks, but is also used for various micromass sensors. Because
the mechanical quality factor Q M is very large (=106), the monitoring resolution of the resonance
frequency reaches Δf R /f R = 10 −6. Thus, even small mass change on the quartz surface can be finely
detected through the resonance frequency shift. Let us examine the mass detection resolution
through the following problem.
EXAMPLE PROBLEM 7.8

A piezoelectric humidity sensor is fabricated using a piezoelectric ceramic plate (5 × 1 × 0.2 mm3
of the structure of Figure 7.14), and driven at its fundamental resonance frequency f R. If humidity
is adsorbed on the piezosensor uniformly, and the resonance frequency is slightly changed by Δf R
(the monitoring resolution of Δf R /f R = 10 −6), evaluate the minimum amount of water adsorbed on
the sensor. Suppose that this frequency change is mainly attributed to the mass/density change,
without changing the elastic constants or sensor size. The piezoceramic density is 8 g/cm3.
Solution
The fundamental resonance frequency is given by
fR = 1 2 L ρs11
E
. (P7.8.1)
Taking “log” for Equation P7.8.1, then their derivatives, we obtain

Bacteria
Antibody/phage: Antibodies
Selectivity Gold
Bacteria Quartz
Substrate Antibodies/phages
(a) (b)
FIGURE 7.25 (a) Principle of a biosensor, where antibody/phage is coated on a single crystal quartz oscil-
lator. Once particular bacteria are captured selectively by the antibodies (b), the surface mass of the oscilla-
tor is increased. (From Cheng, Z.-Y., Private communication, Auburn University, Auburn, AL, 2003. With
permission.)
ΔfR /fR = (1/2)[ Δρ/ρ + Δs11

E E
/s11 ]. (P7.8.2)
Now, Δs11
E /s E = 0 and Δf /f = 10 −6 lead to Δρ/ρ = ΔM/M = 2 × 10 −6.
11 R R
Since M = 8 (g/cm3) × 5 × 1 × 0.2 × 10 −3 = 8 × 10 −3 (g), the detection resolution of the water
weight ΔM will be
ΔM = 8 × 10 −3 × 2 × 10 −6 = 16 × 10 −9 (g) = 16(ng).
This micromass sensor can be utilized for a biosensor for detecting bacteria, such as Escherichia
Coli and Salmonella; 104~107 cells/mL is already critical to humans for Salmonella’s case. Quartz
oscillators can be used to detect this small amount of salmonella bacteria. Figure 7.25a shows the
principle of this biosensor, where antibody/phage is coated on a single crystal quartz oscillator.
Once particular bacteria are captured selectively by the antibodies, the surface mass of the oscillator
is increased (see the SEM photo of captured bacteria in Figure 7.25b). The sensitivity 104 cells/mL
can be obtained.30
7.5.2 VISCOSITY SENSOR

Several companies including Stanford Research Systems, and Ulvac, Japan commercialized a
microbalance to measure the viscosity of the liquid. They used a thickness-shear mode of the AT-cut
quartz for intentionally enhancing the resonance frequency shift by the viscosity.31
7.6 PIEZOELECTRIC TRANSFORMERS

When input and output terminals are fabricated on a piezo-device and input/output voltage is
changed through the vibration energy transfer, the device is called a piezoelectric transformer.
Piezoelectric transformers were used in color TVs in early 1970s, because of their compact size in
comparison with the conventional electromagnetic coil-type transformers. Since serious problems
were found initially in the mechanical strength (fracture occurred at the nodal point!) and in heat
generation, the development approach was the same as that used for fabricating ceramic actuators.
Recent laptop computers with a liquid crystal display require a very thin, no electromagnetic-noise
transformer to start the glow of a fluorescent back-lamp. This application has accelerated the devel-
opment of the piezotransformer, and the previous problems (mechanical strength and heat genera-
tion) have almost been overcome.
Since the original piezo-transformer was proposed by C. A. Rosen32, there have been a variety
of such transformers investigated. Figure 7.26 shows a fundamental structure where two differ-
ently poled parts coexist in one piezoelectric plate. A standing wave with a wavelength equal to the

L1 L2
Low-voltage
input w
t High-voltage
output
FIGURE 7.26 Piezoelectric transformer proposed by Rosen. (From Rosen, C.A., Proc. Electr. Component
Symp., 205, 1957.)
l2
l1
P
P
t
P Output
P
P
Input
w
FIGURE 7.27 ML-type transformer by NEC. (From Kawashima, S. et al., Proc. IEEE Int’l Ultrasonic
Symp. ‘94, France, Nov. 1994.)
sample length is excited, a half wavelength existing on both the input (L1) and output (L2) parts. The
voltage rise ratio r (step-up ratio) is given for the unloaded condition (open circuit) by
( )
r = 4 π 2 k31k33QM (L2 t )⎡2 s33
⎢⎣
E E
s11 1 + s33
D E ⎤
s11
⎥⎦(. ) (7.44)
The r ratio is increased with an increase of (L2/t), where t is the thickness.

NEC proposed a ML-type transformer (Figure 7.27) in order to increase the voltage rise ratio.33
Usage of the third-order longitudinal mode is another idea to distribute the stress concentration.
EXAMPLE PROBLEM 7.9

Using Mason’s equivalent circuits for two length expander bars, surface and end electroded, as
shown in Figure 7.28, calculate the step-up voltage ratio for a Rosen-type transformer under an
open-circuit condition. (The result must be Equation 7.44.)
Solution
A complete equivalent circuit for a length expander bar (top and bottom surface electroded) with
electric field perpendicular to the direction of wave propagation is provided in Figure 7.28a,
where
Z1i = Z 0i tanh ( jωL /2vbE ), (P7.9.1)
Z 2i = Z 0i /sinh ( jωL /vbE ). (P7.9.2)

+
V
–
– V +
e1 Z1 Z1 e2 e1 Z1 Z1 e2
Z2 Z2
F1 1:N F1 –C0 N:1
F2 F2
V V
C0
C0
(a) (b)
FIGURE 7.28 Mason’s equivalent circuits for length expander bars: (a) surface electroded and (b) end
electroded.
L, w, and t are the length, width, and thickness of the sample, vEb is the wave velocity given by
vbE = 1 ρs11
E
. (P7.9.3)
The characteristic mechanical impedance Z 0 and the clamped capacitance C0 are provided by
ρ
Z 0i = ρwtvbE = wt E
, (P7.9.4)
s11
C0i = Lwε 33
X
(1 − k31
2
)/t, (P7.9.5)
and the force factor N is expressed as
N i = wd31 /s11
E
= w(ε 33
X E 1/ 2
/s11 ) k31. (P7.9.6)
In a similar fashion, the necessary parameters for a length expander bar (end electroded) in
Figure 7.28b with electric field parallel to direction of wave propagation are given by
Z1o = Z 0o tanh ( jωL /2vbD ), (P7.9.7)
Z 2o = Z 0o / sinh ( jωL /2vbD ), (P7.9.8)
vbD = l ρs33
D
, (P7.9.9)
ρ
Z 0 o = ρ wt vbD = wt D
, (P7.9.10)
s33
C0o = Lwε 33
X
(1 − k33
2
)/L, (P7.9.11)

l 1:N Z2 Z1 l –C0 1:N Z2 Z1
C0 C0 Z1
V Z1 V
(a) (b)
l1 1:Ni Zi2 Zi1 Zo1 Zo2 N0:1 –C00 l2
V1 Ci0 Zi1 Z01 V2

C00
(c)
FIGURE 7.29 Equivalent circuits for (a) one-end free length expander bar (surface electroded), (b) one-end
free length expander bar (end electroded), and (c) the Rosen-type transformer.
N o = wtd33 /Ls33
D
= (wt /L )(ε 33
X D 1/ 2
/s33 ) k33 . (P7.9.12)
When one end of the piezoelectric element is free (F1 = 0), the above two equivalent circuits
become Figure 7.29a and b. Thus, connecting the F2 sides of the two circuits, we obtain the
equivalent circuit for the Rosen-type transformer (1st mode application) as shown in Figure
7.29c. Note that L in all the equations, Equations. P7.9.1 through P7.9.12 must be replaced by L1
and L2 for the transformer configuration shown in Figure 7.26.
The step-up voltage ratio for an open-circuit condition can be estimated by
r = 2 N i N o /ωC0o R, (P7.9.13)
where
C0o = wt ε 33
X
(1 − k33
2
)/L2 , (P7.9.14)
R = πZ 0o /QM , (P7.9.15)
(
ω = 2π vbD 2L2 = π L2 ) ρ
D
s33
. (P7.9.16)
Thus, taking into account the relation
D
s33 E
/s33 = 1 − k33
2
, (P7.9.17)
r = 2 ⎡ w ε 33( ) (
k31 ⎤ ⎡(wt L2 ) ε 33 ) k33 ⎤
12 12
X E X D
s11 s11
⎣⎢ ⎦⎥ ⎣⎢ ⎦⎥
⎡ ρ ⎤ ⎡ ⎤
( ) ρ
⎢ π L2 D ⎥ ⎡⎣ wt ε 33 1 − k33 L2 ⎤⎦ ⎢ π wt D QM ⎥
X 2
⎣⎢ s33 ⎥
⎦ ⎢⎣ s33 ⎥⎦
( )
= 4 π 2 k31k33QM (L2 t )⎡2 s33
⎣⎢
E E
s11 (1 + D
s33 E ⎤
s11 )
⎦⎥
. (P7.9.18)
Step-down transformers for adaptor applications for portable equipment such as laptop comput-
ers and mobile phones have also been developed. Figure 7.30 shows a credit-card size 35 W
adaptor for a laptop computer, developed by Face Electronics, Taiheiyo Cement, in collaboration
with The Penn State University.34

FIGURE 7.30 Credit-card size laptop computer adaptor (35 W) using a piezoelectric step-down transformer
(below), in comparison with a commercial adaptor with an electromagnetic transformer (top).
7.7 PIEZOELECTRIC ACTUATORS

Piezoelectric and electrostrictive devices have become key components in smart actuator systems
such as precision positioners, miniature ultrasonic motors (USMs), and adaptive mechanical dampers.
This section reviews the developments of piezoelectric and related ceramic actuators with particular
focus on the improvement of actuator materials, device designs, and applications of the actuators.
Piezoelectric actuators are forming a new field between electronic and structural ceramics.35–38
Application fields are classified into three categories: positioners, motors, and vibration suppressors.
The manufacturing precision of optical instruments such as lasers and cameras, and the positioning
accuracy for fabricating semiconductor chips, which must be adjusted using solid-state actuators,
are generally on the order of 0.1 μm. Regarding conventional electromagnetic motors, tiny motors
smaller than 1 cm are often required in office or factory automation equipment and are rather dif-
ficult to produce with sufficient energy efficiency. USMs whose efficiency is insensitive to size are
considered superior in the minimotor area. Vibration suppression in space structures and military
vehicles using piezoelectric actuators is another promising field of application.
New solid-state displacement transducers controlled by temperature (shape memory alloy) or
magnetic field (magnetostrictive alloy) have been proposed, but are generally inferior to the piezo-
electric/electrostrictive ceramic actuators because of current technological trends aimed at reduced
driving power and miniaturization.38 The shape memory actuator is too slow in response with a
very low energy efficiency, while the magnetostrictor requires a driving coil which is very bulky
and generates magnetic noise.
7.7.1 CERAMIC ACTUATOR MATERIALS

Actuator materials are classified into three categories: piezoelectric, electrostrictive, and phase-
change materials. Modified PZT [Pb(Zr,Ti)O3] ceramics are currently the leading materials for

piezoelectric applications. The PLZT [(Pb,La)(Zr,Ti)O3] 7/62/38 compound is one such composi-
tion.39 The strain curve is shown in Figure 7.31a left. When the applied field is small, the induced
strain x is nearly proportional to the field E (x = dE, where d is called the piezoelectric constant). As
the field becomes larger (i.e., greater than about 1 kV/cm), however, the strain curve deviates from
this linear trend and significant hysteresis is observed due to polarization reorientation. This some-
times limits the use of such materials for actuator applications that require nonhysteretic response.
An interesting new family of actuators has been fabricated in Germany from the barium stannate
titanate system [Ba(Sn,Ti)O3].40 The useful property of Ba(Sn0.15Ti0.85)O3 is its unusual strain curve,
in which the domain reorientation occurs only at low fields, and there is then a long linear range at
higher fields [Figure 7.31a right]; that is, the coercive field is unusually small. Moreover, this system
is particularly intriguing since it contains no Pb ions, an essential feature as ecological concerns
grow nowadays.
The second category of actuators is based on electrostriction as exhibited by PMN [Pb(Mg1/3Nb2/3)
O3]-based ceramics. Although it is a second-order phenomenon of electromechanical coupling
(x = ME2, where M is called the electrostrictive coefficient), the induced strain can be extraordinarily
large (more than 0.1%).41 An attractive feature of these materials is the near absence of hysteresis
(Figure 7.31b). The superiority of PMN to PZT was demonstrated in a scanning tunneling micro-
scope (STM).42 The STM probe was scanned mechanically by a PMN actuator, which produces
only extremely small distortion of the image, even when the probe was scanned in the opposite
direction due to a negligibly small hysteresis.
The third category is based on phase-change-related strains, that is, polarization induced by
switching from an antiferroelectric to a ferroelectric state.43 Figure 7.31c shows the field-induced
strain curves taken for the lead zirconate stannate-based system [Pb0.99Nb0.02( (ZrxSn1−x)1−yTiy)0.98O3].
The longitudinally induced strain reaches more than 0.3%, which is much larger than that exhibited
Piezostrictor Electrostrictor
PLZT BST PMN–PT
Strain
4 ×10–4 1 ×10–3
Strain
Strain
6
×10–3
3 5 0.75
4
2 0.5
3
2
1
1
–15 –10 –5 0 5 10 15 –20 –10 0 10 20 –15 –10 –5 0 5 10 15
(a) Electric field (kV/cm) Electric field (kV/cm) (b) Electric field (kV/cm)
Phase-change material
PNZST PNZST
Strain (×10–3)
3 3
Strain (×10–3)
2 2
1 1
–30 –20 –10 0 10 20 30 –30 –20 –10 0 10 20 30

(c) Electric field (kV/cm) Electric field (kV/cm)
FIGURE 7.31 Electric field-induced strains in various ceramics: (a) piezoelectric (Pb,La)(Zr,Ti)O3, Ba(Sn,Ti)
O3, (b) electrostrictive Pb(Mg1/3Nb2/3)O3, and (c) phase-change materials Pb(Zr,Sn,Ti)O3.

Moving piece
Leaf spring
Contact
Unimorph
Base
10 mm
FIGURE 7.32 Latching relay using a shape memory ceramic unimorph. The drive requires a 4 ms pulse volt-
age, not a continuous voltage, which produces a 150 μm unimorph tip displacement.
by normal piezostrictors or electrostrictors. A rectangular-shaped hysteresis in Figure 7.31c left char-

acterizes the response of these devices which are referred to as “digital displacement transducers”
because of the two on/off strain states. Moreover, this field-induced transition is accompanied by a
shape memory effect for appropriate compositions (Figure 7.31c right). Once the ferroelectric phase
has been induced, the material “memorizes” its ferroelectric state even under zero-field conditions,
although it can be erased with the application of a small reverse bias field.44 This shape memory
ceramic is used in energy-saving actuators. A latching relay in Figure 7.32 is composed of a shape
memory ceramic unimorph and a mechanical snap action switch, which is driven by a pulse voltage
of 4 ms duration. Compared with the conventional electromagnetic relays, the new relay is much
simpler and more compact in structure with almost the same response time.
7.7.2 ACTUATOR DESIGNS

Two of the most popular actuator designs are the MLs45 and bimorphs (see Figure 7.33). The ML,
in which roughly 100 thin piezoelectric ceramic sheets are stacked together, has the advantages of
low driving voltage (100 V), quick response (10 μs), high generative force (1 kN), and high electro-
mechanical coupling. But the displacement, on the order of 10 μm, is not sufficient for some applica-
tions. This contrasts with the characteristics of the bimorph which consists of multiple piezoelectric
and elastic plates bonded together to generate a large bending displacement of several hundred
micrometers, but has relatively low response time (1 ms) and generative force (1 N).
A 3-D positioning actuator with a stacked structure as pictured in Figure 7.34 was proposed by a
German company, where shear strain was utilized to generate the x and y displacements.46 Polymer-
packed PZT bimorphs have been commercialized by ACX for vibration reduction/control applica-
tions in smart structures.47
The market research conducted in 1998 by Japan Technology Transfer Association clarified that
the actual demands on the actuators are as follows: 100 μm displacement, 100 N force, and 100 μs
response. Because neither the ML nor the bimorph can satisfy the actual demand, composite actua-
tor structures called “moonie” and “cymbal” have been developed to provide characteristics inter-
mediate between the ML and bimorph actuators; this transducer exhibits an order of magnitude
larger displacement than the ML, and much larger generative force with quicker response than the
bimorph.48 The device consists of a thin ML piezoelectric element and two metal plates with a nar-
row moon-shaped or cymbal-shaped cavity bonded together as shown in Figure 7.33. The moonie
with a size of 5 × 5 × 2.5 mm3 can generate a 20 μm displacement under 60 V, 8 times as large as the
generative displacement produced by a ML of the same size.49 This new compact actuator has been
utilized in a miniaturized laser beam scanner.

Multilayer
Single plate Moonie
Bimorph
FIGURE 7.33 Typical designs for ceramic actuators: ML, moonie, and bimorph.
Z-stack (10 layers)

(extension)
X-stack (10 layers)

(shear)
Y-stack (10 layers)

(shear)
FIGURE 7.34 3-D positioning ML actuator. Notice that the x- and y-stacks are using shear mode with the
spontaneous polarization perpendicular to the applied electric field direction.
7.7.3 DRIVE/CONTROL TECHNIQUES

Piezoelectric/electrostrictive actuators may be classified into two categories, based on the type of
driving voltage applied to the device and the nature of the strain induced by the voltage (Figure 7.35):
(1) rigid displacement devices for which the strain is induced unidirectionally along the direction
of the applied DC field and (2) resonating displacement devices for which the alternating strain

Servo displacement
E Feedback x transducer
Eb
Servo
drive Electrostrictive material
(Hysteresis-free)
Rigid t Eb E
strain Em Pulse x
E ON Pules drive motor
On/Off
drive Soft piezoelectric
material
t OFF Em E (Low permittivity)
E x
Ultrasonic motor
Resonant AC Sine
strain drive Hard piezoelectric
t E material
(High Q)
FIGURE 7.35 Classification of piezoelectric/electrostrictive actuators.
is excited by an AC field at the mechanical resonance frequency (USMs). The first can be further
divided into two types: servo displacement transducers (positioners) controlled by a feedback system
through a position-detection signal, and pulse-drive motors operated in a simple on/off switching
mode, exemplified by ink-jet printers.
The material requirements for these classes of devices are somewhat different, and certain com-
pounds will be better suited to particular applications. The USM, for instance, requires a very hard
piezoelectric with a high mechanical quality factor Q M, to suppress heat generation. Note that the
resonating strain/displacement is amplified by a factor of Q M, in comparison with the off-resonance
strain/displacement (i.e., dEL).50 Driving the motor at the antiresonance frequency, rather than at
resonance, is also an intriguing technique to reduce the load on the piezoceramic and the power sup-
ply.51 The servo displacement transducer suffers most from strain hysteresis and, therefore, a PMN
electrostrictor is used for this purpose. The pulse-drive motor requires a low permittivity material
aimed at quick response with a certain power supply (a high-power supply is expensive from the
practical device application viewpoint) rather than a small hysteresis, so soft PZT piezoelectrics are
preferred rather than the high-permittivity PMN for this application.
Pulse-drive techniques for ceramic actuators are very important for improving the response of
the device.52,53 Figure 7.36 shows transient vibrations of a bimorph excited after a pseudostep volt-
age is applied. The rise time is varied around the resonance period (n is the timescale with a unit of
T0/2, where T0 stands for the resonance period). It is concluded that the overshoot and ringing of the
tip displacement is completely suppressed when the rise time is precisely adjusted to the resonance
period of the piezodevice (i.e., for n = 2).52 A flight actuator was developed using a pulse-drive
piezoelectric element and a steel ball. A 5 μm rapid displacement induced in a ML actuator can hit a
2 mm steel ball up to 20 mm in height. A dot-matrix printer head has been developed using a flight
actuator as shown in Figure 7.37.54 By changing the drive voltage pulse width, the movement of the
armature was easily controlled to realize no vibrational ringing or double hitting.

Consider the longitudinal mechanical vibration in a piezoceramic plate with thickness b, width
w, and length L (b << w << L), as shown in Figure 7.14. If the polarization is in the z-direction and
the x–y planes are the planes of the electrodes, the extensional vibration in the x-direction is
represented by the following dynamic equation:
ρ (∂ 2u /∂t 2 ) = F = (∂X11 /∂x ) + (∂X12 /∂y) + (∂X13 /∂z), (P7.10.1)

Tip
displacement
Electric field
(a) n = 1
Tip
displacement
Electric field
10 ms
(b) n = 2
Tip
displacement
Electric field
(c) n = 3
FIGURE 7.36 Transient vibration of a bimorph excited after a pseudostep voltage applied. Here, n is a
timescale with a unit of half of the resonance period (i.e., 2n = the resonance period).
Ink ribbon
Armature
Steel plate
Piezoelectric Wire
actuator
Spring
Platen
Paper
10 mm
Base
FIGURE 7.37 Dot-matrix printer head using a flight actuator mechanism.
where u is the displacement of a small volume element in the ceramic plate in the x-direction.
When the plate is very long and thin, the longitudinal stresses, X2 and X3 may be set equal to
zero throughout the plate, and shear stress is not generated by the electric field Ez, the necessary
piezoelectric equation is reduced to
X1 = x1 /s11
E
− (d31 /s11
E
)Ez (P7.10.2)
Thus, we obtain the following dynamic equation (note that this equation does not include a loss
during vibration):
ρ (∂ 2u /∂t 2 ) = (1/s11
E
) (∂ 2u /∂x 2 ). (P7.10.3)

Provide the displacement response to a pseudostep voltage, as demonstrated in Figure 7.36.
Solution
Let us solve the differential equation (Equation P7.10.3), using the Laplace transform. Denoting
~
the Laplace transforms of u(t, x) and Ez(t) as U(s, x) and E(s), respectively, Equation P7.10.3 is
transformed to
ρs11s U (s, x ) = ∂ 2U (s, x )/∂x 2 .

E 2
(P7.10.4)
This assumes the following initial conditions:
u(t = 0, x ) = 0, ∂u(t = 0, x )/∂t = 0. (P7.10.5)
Using
ρs11
E
= 1/v 2 , (v: sound velocity in the piezoceramic) (P7.10.6)
we obtain a general solution
U (s, x ) = Ae( sx / v ) + Be − ( sx / v ) . (P7.10.7)
The constants A and B can be determined by the boundary conditions X1 = 0 at x = 0 and L. Note
that the strain x1 at x = 0 and L may not need to be zero:
X1 = ( x1 − d31Ez )/s11
E
= 0. (P7.10.8)
Using the fact that the strain x1 = ¶u/¶x
L[ x1 ] = ∂U /∂x = A(s /v)e( sx / v ) − B(s /v)e − ( sx / v ) , (P7.10.9)
the boundary conditions provide the relations:
A (s /v ) − B (s /v ) = d31E,
A (s /v )e(sL /v ) − B (s /v )e − (sL /v ) = d31E. (P7.10.10)
Thus, we obtain
(
A = d31E 1 − e ( )
− sL /v
) (s /v)(e (sL /v )
)
− e − (sL /v ) , (P7.10.11)
(
B = d31E 1 − e(sL /v ) ) (s /v)(e (sL /v )
)
− e − (sL /v ) , (P7.10.12)
and consequently, Equations P7.10.7 and P7.10.9 become
(
U (s,x ) = d31E(v /s) ⎡ e − s (L − x )/v + e − s (L + x )/v − e − sx /v − e − s ( 2L − x )/v
⎣ ) (1 − e −2sL /v
)⎤⎦ , (P7.10.13)
(
L[x1 ] = d31E ⎡ e − s (L − x )/v − e − s (L + x )/v + e − sx /v − e − s (2L − x )/v
⎣ ) (1 − e −2sL /v
)⎤⎦ . (P7.10.14)
The inverse Laplace transforms of Equations P7.10.13 and P7.10.14 now provide the displace-
ment u(t, x) and strain x1(t, x). Since u(t, x = L/2) = 0 [from U(s, x = L/2) = 0] and u(t, x = 0) = −u(t,
x = L) [from U(s, x = 0) = −U(s, x = L)], the total displacement of the plate device ΔL becomes
equal to 2u(t, x = L).

(
U (s,x = L ) = d31E(v /s) 1 − e − (sL /v ) ) (1 + e − (sL /v )
)
= d31E(v /s)tanh(sL /2v). (P7.10.15)
~
When E includes the term (1 + e−(sL/v)), the denominator of Equation P7.10.15 disappears and the
~
displacement u(t, x) is not accompanied by the vibration ringing. However, when E does not include
the term of (1 + e−(sL/v)), taking into account the expansion series
1/(1+ e − sL / v ) = 1− e − sL / v + e −2sL / v − e −3sL / v ... (P7.10.16)
the strain along the x direction, x1(t, x) will be obtained by superimposing the d31Ez(t) curves with
a specific time delay, and the ringing of the displacement will continue forever.
Let us consider the displacement response to a pseudostep voltage shown in Figure 7.38).
Substituting
E = (E0 v /nLs 2 )(1 − e − snL / v ) (P7.10.17)
into Equation P7.10.16, let us obtain the displacement ΔL for n = 1, 2, and 3.
1. n = 1:
U (s, L ) = (d31E0 /L )(v 2 /s 3 )(1 − e − sL / v )2 /(1 + e − sL / v )
= (d31E0 /L )(v 2 /s 3 )(1 − 3e − sL / v + 4e −2 sL / v − 4e −3sL / v + . (P7.10.18)
Note that the base function of U(s, L), 1/s3, gives the base function of u(t, L) as t2/2 (parabolic
curve). Thus, considering that the exponential term in the Laplace s domain corresponds to the
ΔL n=1
E(t) Resonance period:
T = 2I/v (3/2)d31E0L
E0
d31E0L
(1/2)d31E0L
0 nI/v t
0 T 2T t
ΔL n=2 ΔL n=3
(7/6)d31E0L
d31E0L d31E0L
(5/6)d31E0L
0 T 2T t 0 T 2T t
FIGURE 7.38 Transient displacement induced in a rectangular plate for a pseudostep voltage. The time scale
n is based on half of the resonance period T.

time shift in the real time domain [for example, e−2sL/v in Equation P7.10.18 corresponds to the
time shift of 2L/v of the base u(t, L) curve t2, that is, (t −2L/v)2]
u(t, L ) = (d31E0 v 2 /2 L )t 2 [0 < t < L /v]
u(t, L ) = (d31E0 v 2 /2 L ) [t 2 − 3(t − L /v)2 ]
= (d31E0 v 2 /2 L ) [ −2(t − (3/2)L /v)2 + (3/2)L2 /v 2 ] [ L /v < t < 2 L /v]
u(t, L ) = (d31E0 v 2 /2 L ) [t 2 − 3(t − L /v)2 + 4(t − 2 L /v)2 ]
= (d31E0 v 2 /2 L ) [2(t − (5/2)L /v)2 + (1 / 2)L2 /v 2 ] [2 L /v < t < 3L /v]
u(t, L ) = (d31E0 v 2 /2 L )[t 2 − 3(t − L /v)2 + 4(t − 2 L /v)2 − 4(t − 3L /v)2 ] [3L /v < t < 4 L /v].
(P7.10.19)
Figure 7.38a shows the change of ΔL (=2u(t, L) ) with time, which exhibits continuous ringing.
Notice that this curve is a combination of parabolic curves, not a sinusoidal curve. The height of
the overshoot is 1/2 of d31E 0 L.
2. n = 2:
U (s, L ) = (d31E0 /2 L )(v 2 /s 3 )(1 − 2e − sL / v + e −2

sL / v
). (P7.10.20)
Thus
u(t, L ) = (d31E0 v 2 /4 L )t 2 [0 < t < L /v]
u(t, L ) = (d31E0 v 2 /4 L )[t 2 − 2(t − L /v)2 ]
= (d31E0 v 2 /4 L )[ −(t 2 − 2 L /v)2 + 2 L2 /v 2 ] [ L /v < t < 2 L /v]
u(t, L ) = (d31E0 v 2 /4 L )[t 2 − 2(t − L /v)2 + (t − 2 L /v)2 ]
= d31E0 L /2 [2 L /v < t ]. (P7.10.21)

~
ΔL does not exhibit ringing, as shown in Figure 7.38b. When the applied field E includes the term
(1 + e−sL/v), the expansion series terminates in finite terms, leading to a complete suppression of
vibrational ringing.
3. n = 3:
U(s, L) is again expanded as an infinite series:
U (s, L ) = (d31E0 /3L )(v 2 /s 3 )(1 − e − sL / v )(1 + e −3sL / v )/(1 + e − sL / v )
= (d31E0 /3L )(v 2 /s 3 )(1 − 2e − sL / v + 2e −2 sL / v − 3e −3sL / v + 4e −4 sL / v − 4e −5sL / v

(P7.10.22)
Figure 7.38c shows the displacement change with time. Note again that all the curves are
composed of parabolic curves and that the height of the overshoot is 1/6 of d31E 0 L.
For n = 4, you can eliminate the ringing again. This derivation suggests an empirical process
on how to suppress the overshoot and ringing. (1) By applying a relative steep rising voltage to the
actuator, we can obtain the resonance period from the time period between the overshoot peak
and the successive peak point. (2) By adjusting the voltage rise time exactly to the resonance
period next time, we can eliminate the overshoot and ringing of the vibration.

7.7.4 DEVICE APPLICATIONS

7.7.4.1 Servo Displacement Transducers
A typical example is found in a space truss structure proposed by the Jet Propulsion Laboratory.55
A stacked PMN actuator was installed at each truss nodal point and operated so that unnecessary
mechanical vibration was suppressed immediately. A “hubble” telescope has also been proposed
using ML PMN electrostrictive actuators to control the phase of the incident light wave in the field
of optical information processing (Figure 7.39).56 The PMN electrostrictor provided superior adjust-
ment of the telescope image because of negligible strain hysteresis.
The U.S. Army is interested in developing a rotor control system in helicopters. Figure 7.40 shows
a bearingless rotor flexbeam with attached piezoelectric strips.57 Various types of PZT-sandwiched
beam structures have been investigated for such a flexbeam application and for active vibration
control.58
Concerning home appliance applications, there was a large market in VCR systems previously.
The requirement for high-quality images has become very stringent for VCRs especially when
played in still, slow or quick mode. As illustrated in Figure 7.41, when the tape is running at a speed
different from the normal speed, the head trace deviates from the recording track depending on the
Lightweighted Bezel Tilt

mirror retainers Bezel mechanism Cover
Tie bar Pin flexures PMN actuators

flexures
FIGURE 7.39 “Hubble” telescope using three PMN electrostrictive ML actuators for optical image correction.
Lag pin
Piezoelectric Torque tube

crystals
Hub
Flexbeam
Pitch link
Blade
FIGURE 7.40 Bearingless rotor flexbeam with attached piezoelectric strips. A slight change in the blade
angle provides enhanced controllability.

• Normal speed head locus

Double speed
• Double speed (with bimorph function) (without bimorph
function)
Tape direction
25.35 mm
Video
track
Tape
130 μm Track width

guard band
(a) 51 μm
(b) Head bimorph
FIGURE 7.41 Locus of the video head and the function of the piezoactuator.
velocity difference. Thus the head traces on the guard band, generating guard band noise.59 The
autotracking scan system by Ampex operates with a piezoelectric actuator so that the head follows
the recording track. The piezoelectric device generates no magnetic noise.
Bimorph structures are commonly used for this tracking actuator because of their large displace-
ment. However, special care has been taken not to produce a spacing angle between the head and
the tape, because a single bimorph exhibits deflection with slight rotation. Figure 7.42 summarizes
various designs proposed to produce a completely parallel motion of the bimorphs.
Figure 7.42a is the Ampex’s S-shape mode bimorph, in which the tip of the bimorph bends oppo-
sitely to cancel the rotation angle. In order to detect the tape track, the head is vibrated slightly at
450 Hz, and this modulation signal is monitored to be fed-back to the actuator. The sensor electrode
is for monitoring the head position, as well as for electrical damping.
Figure 7.42b shows a parallel spring type developed by SONY.60 This actuator was installed on a
broadcasting VCR, BVH-1000 (Figure 7.41b). Since the track width is wide (130 μm) for obtaining
high-resolution images, a large displacement is required for the actuator.
Figure 7.42c shows a ring-type bimorph developed for consumer use by Matsushita Electric.61,62
Using this compact movable head, noiseless picture search has become possible both in quick scan
(up to 8 times) and still/slow mode. Autotracking using a piezoelectric actuator was considered as a
standard device in 8 mm VCRs. The reader can adopt the above three mechanisms in order to obtain
the perfect parallel motion with bimorphs.
7.7.4.2 Pulse-Drive Motors

A dot-matrix printer is the first widely commercialized product using ceramic actuators. Each char-
acter formed by such a printer was originally composed of a 24 × 24 dot matrix. A printing ribbon is
subsequently impacted by a multiwire array. A sketch of the printer head is given in Figure 7.43a.63 The
printing element is composed of a ML piezoelectric device, in which 100 thin ceramic sheets 100 μm
in thickness were stacked, together with a sophisticated magnification mechanism (Figure 7.43b).
The magnification unit is based on a monolithic hinge lever with a magnification of 30, resulting in
an amplified displacement of 0.5 mm and energy transfer efficiency greater than 50%.
Figure 7.44 illustrates the recent ink-jet printer produced by Seiko Epson,64 in which PZT thin
plates were laminated with vibration, chamber, and communication plates to create a unimorph
actuation mechanism. Using the cofiring technique with PZT and ZrO2 elastic parts for manufac-
turing this ML chips head (MACH), Epson achieved superior stability in the ink chamber vibration
and for various inks. In addition, more importantly, the manufacturing cost reduced dramatically by
adopting this cofiring technique.

Subbimorph
Mother bimorph
Head holder
Sensor electrode
L2 Head chip
L1
Support
L L3
Actuator electrode
(a) (b)
Piezo device support
Video head D
Piezoelectric device
Video head C
Piezo device fixer
Video head B
Video head A
(Record/play)
(Record/play)
Rotary cylinder
(c)
FIGURE 7.42 VCR head tracking actuators: (a) complex bimorph by Ampex, (b) parallel bimorph by SONY,
and (c) ring bimorph by Matsushita Electric.
Head
element
Platen
Paper
Ink ribbon
Guide
Piezoelectirc
actuator
Wire
Wire Stroke amplifier

Wire guide
(a) (b)
FIGURE 7.43 (a) Structure of a dot-matrix printer head (NEC), and (b) a differential-type piezoelectric printer-
head element. A sophisticated monolithic hinge lever mechanism amplifies the actuator displacement by 30 times.

ML Chips
Upper electrode
PZT
Lower electrode 0.5 mm
Vibration plate
Chamber plate
Communication plate
Adhesive layer Ink pass
Stainless plate Ink supply hole

Ink chamber
Nozzle
FIGURE 7.44 MACH ink-jet printer head developed by Seiko Epson. (From Kurashima, N., Proc. Machine
Tech. Inst. Seminar, MITI, Tsukuba, Japan, 1999. With permission.)
Toyota developed a Piezo Toyota Electronic Modulated Suspension (TEMS), which responds to
each protrusion on the road by adjusting the damping condition. TEMS was installed on a “Celcio”
(equivalent to Lexus, internationally) in 1989.65 In general, as the damping force of a shock absorber
in an automobile is increased (i.e., “hard” damper), the controllability and stability of a vehicle are
improved. However, comfort is sacrificed because the road roughness is easily transferred to the
passengers. The purpose of the electronically controlled shock absorber is to obtain both controlla-
bility and comfort simultaneously. Usually the system is set to provide a low damping force (“soft”)
so as to improve comfort and the damping force is changed to a high position according to the road
condition and the car speed to improve the controllability. In order to respond to a road protrusion,
a very high response of the sensor and actuator combination is required.
Figure 7.45 shows the structure of the electronically controlled shock absorber. The sensor is
composed of five layers of 0.5 mm thick PZT disks. The detecting speed of the road roughness
is about 2 ms and the resolution of the up-down deviation is 2 mm. The actuator is made of 88
layers of 0.5 mm thick disks. Applying 500 V generates a displacement of about 50 μm, which
Piezoelectric sensor
Piezoelectric multilayer
actuator
Piston
Damping change
valve
FIGURE 7.45 TEMS with a ML piezoelectric actuator and a sensor.

1s
Automatic control
Up-down acceleration 0.1 m/s2
Pitching rate 2°/s
Fixed at hard damping

Up-down acceleration
Pitching rate
Fixed at soft damping

Up-down acceleration
Pitching rate
FIGURE 7.46 Function of the TEMS-adaptive suspension system. Note that both up-down acceleration and
pitching rate are minimized under the TEMS system.
is magnified by 40 times through a piston and plunger pin combination. This stroke pushes the
change valve of the damping force down, then opens the bypass oil route, leading to the decrease
of the flow resistance (i.e., “soft”).
Figure 7.46 illustrates the operation of the suspension system. The up-down acceleration and
pitching rate were monitored when the vehicle was driven on a rough road. When the TEMS system
was used (top figure), the up-down acceleration was suppressed to as small as the condition fixed at
soft, providing comfort. At the same time, the pitching rate was also suppressed to as small as the
condition fixed at “hard,” leading to better controllability.
In order to increase the diesel engine efficiency, high-pressure fuel and quick injection control
are required. Figure 7.47 shows an example of diesel fuel injection timing chart. In this one cycle
(typically 60 Hz), the multiple injections should be realized in a very sharp shape. For this pur-
pose, piezoelectric actuators, specifically ML types, were adopted. The highest reliability of these
devices at an elevated temperature (150°C) for a long period (10 years) has been achieved. So-called
common-rail-type injection valves have been widely commercialized by Siemens, Bosch, and Denso
Corporation (Figure 7.48).66
Figure 7.49 shows a walking piezomotor with four ML actuators developed by Philips.67 Two
shorter actuators function as clamps and the longer two provide the movement by an inchworm
mechanism. A major drawback of this inchworm design is the mechanical noise created by the
on–off drive (audible frequency due to the requirement lower than the mechanical resonance of the
system).
Main
Pilot Pre After Post
Injection interval
FIGURE 7.47 Diesel fuel injection timing chart in one cycle (about 60 Hz cycle). (From Fujii, A., Proc.
JTTAS Meeting on Dec. 2, Tokyo, 2005. With permission.)

Piezoelectric actuator
Control valve
Displacement
amplification unit
Injector body
Nozzle
FIGURE 7.48 Common rail-type diesel injector with a piezoelectric ML actuator. (From Fujii, A., Proc.
33
Multilayer piezo actuator Fv
30 4
d33
θ2.5
A B
Fo d33
Fv
d31
d31
Fo
Multilayer piezo actuator
FIGURE 7.49 Walking piezomotor using an inchworm mechanism with four ML piezoelectric actuators by
Philips.
7.8 ULTRASONIC MOTORS

7.8.1 ULTRASONIC MOTORS (USM)
The USM is one of the piezoelectric actuator categories. However, since it has already been matur-
ing as an industry, the author uses a different section for its discussion.
Electromagnetic motors were invented more than 100 years ago. While these motors still domi-
nate the industry, a drastic improvement cannot be expected except through new discoveries in mag-
netic or superconducting materials. Regarding conventional electromagnetic motors, tiny motors
smaller than 1 cm are rather difficult to produce sufficient energy efficiency (more than 90% input
electrical energy is spent for heat generation in a wristwatch motor at present). Therefore, a new class
of motors using high-power ultrasonic energy, i.e., USM, is gaining widespread attention. USMs
made with piezoceramic whose efficiency is insensitive to size are superior in the minimotor area.
Figure 7.50 shows the basic construction of most USMs, which consist of a high-frequency power
supply, a vibrator, and a slider. The vibrator is composed of a piezoelectric driving component and
an elastic vibratory part. The slider is composed of an elastic moving part and a friction coat.

Stator
Piezoelectric Elastic vibrator

driver piece
Electrical
input Mechanical
output
High-frequency
power supply
Friction Elastic sliding
cost piece
Slider/rotor
FIGURE 7.50 Fundamental construction of an USM.
Rotor Horn 2
Horn 1
Piezo vibrator 2
Piezo vibrator 1
FIGURE 7.51 USM by Barth. (From Barth, H.V., IBM Techn. Discl. Bull., 16, 2263, 1973.)
Although there had been some earlier attempts, the first practical USM was proposed by H. V.
Barth of IBM in 1973.68 As shown in Figure 7.51, the rotor was pressed against two horns placed
at different locations. By exciting one of the horns, the rotor was driven in one direction, and by
exciting the other horn, the rotation direction was reversed. Various mechanisms based on virtually
the same principle were proposed by V. V. Lavrinenko69 and P. E. Vasiliev70 in the former USSR.
Because of difficulty in maintaining constant vibration amplitude with temperature rise, wear, and
tear, the motors were not of much practical use at that time.
In the 1980s, with increasing chip pattern density, the semiconductor industry began to demand
much more precise and sophisticated positioners which would not generate magnetic field noise.
This urgent need accelerated the development of USMs. Another advantage of USMs over conven-
tional electromagnetic motors with expensive copper coils is the improved availability of piezo-
electric ceramics at reasonable cost. Japanese manufacturers are currently producing piezoelectric
buzzers at about ¢30–¢40 per unit.
Let us summarize the merits and demerits of the USM:
Merits
1. Low speed and high torque—Direct drive
2. Quick response, wide velocity range, hard brake, and no backlash
–Excellent controllability
–Fine position resolution

3. High power/weight ratio and high efficiency

4. Quiet drive (driven at an inaudible frequency for human)
5. Compact size and lightweight
6. Simple structure and easy production process
7. Negligible effect from external magnetic or radioactive fields, and also no generation of
these fields
Demerits
8. Necessity for a high-frequency power supply
9. Less durability due to frictional drive
10. Drooping torque vs. speed characteristics
7.8.2 CLASSIFICATION AND PRINCIPLES OF ULTRASONIC MOTORS

From a customer’s point of view, there are rotary- and linear-type motors. If we categorize them
according to the vibrator shape, there are rod-type, π-shaped, ring (square) and cylinder types. Two
categories are being investigated for USMs from a vibration characteristic viewpoint: a standing-
wave type and a propagating-wave type. Refresh your memory on the wave formulas. The standing
wave is expressed by
us ( x, t ) = A cos kx ⋅ cos ωt , (7.45)
while the propagating wave is expressed as
up ( x, t ) = A cos(kx − ωt ). (7.46)
Using a trigonometric relation, Equation 7.46 can be transformed as
U p ( x, t ) = A cos kx ⋅ cos ωt + A cos(kx − π /2) ⋅ cos(ωt − π /2). (7.47)
This leads to an important notion: a propagating wave can be generated by superimposing two
standing waves whose phases differ by 90° both in time and in space. This principle is necessary to
generate a propagating wave on a limited volume/size substance, because only standing waves can
be excited stably in a solid medium of finite size.
7.8.2.1 Standing-Wave Type

The standing-wave type is sometimes referred to as a vibratory-coupler type or a “woodpecker”
type, where a vibratory piece is connected to a piezoelectric driver and the tip portion generates an
elliptical movement. Figure 7.52 shows a simple model proposed by T. Sashida.71 A vibratory piece
is attached to a rotor or a slider with a slight cant angle θ. Take the x–y coordinate so that the x-axis
is normal to the rotor face. When a vibration displacement
ux = u0 sin (ωt + α ), (7.48)
is excited at the piezoelectric vibrator, the vibratory piece generates bending because of restriction
by the rotor, so that the tip moves along the rotor face between A → B, and freely between B → A.
If the vibratory piece and the piezovibrator are tuned properly, they form a resonating structure, and
if the bending deformation is sufficiently small compared with the length, the tip locus during the
free vibration (B → A) is represented by
x = u0 sin (ωt + α),
y = u1 sin (ωt + β), (7.49)

y y
Oscillator Vibratory V0
~ piece M
Rotor B Uk
B'
θ x
x P O FA θ Fr
Fn
δ
A'
γ A
U0 Δ
(a) Piezoelectric vibrator (b)
FIGURE 7.52 Vibratory coupler-type motor (a) and its tip locus (b).
which composes an elliptical locus. Therefore, only the duration A → B provides a unidirectional
force to the rotor through friction, which leads to an intermittent rotational torque or thrust.
However, because of the inertia of the rotor, the rotation speed ripple is not observed to be large.
The standing-wave type, in general, is low in cost (one vibration source) and has high efficiency
(up to 98% theoretically), but lacks control in both the clockwise and counterclockwise directions,
in general.
7.8.2.2 Propagating-Wave Type

By comparison, the propagating-wave type (a surface-wave or “surfing” type) combines two stand-
ing waves with a 90° phase difference both in time and in space. The principle is illustrated in
Figure 7.53. A surface particle of the elastic body draws an elliptical locus due to the coupling of
longitudinal and transverse waves. This type requires, in general, two vibration sources to generate
one propagating wave, leading to low efficiency (not more than 50%), but it is controllable in both
rotational directions.

The thrust/torque of the USM is obtained by the friction force. Referring to Figure 7.53 (traveling
wave–type motor), consider the model to calculate the linear motor speed and thrust.
z
Slider
Moving direction
A A'
x
w Propagation direction
u
Elastic body
FIGURE 7.53 Principle of the propagating wave–type motor.

Solution
Suppose that the slider is pressed on the motor stator with the force N, we can calculate the
maximum friction force (without slip on the surface) by
F = μN , (P7.11.1)
where μ is the static friction constant, and around 0.2 for the metal–metal contact, and 0.4 for
the metal–polymer contact. For example, when N = 10 N (almost equivalent to 1 kg weight of the
slider), 2 N thrust can be expected. When the load is higher than this level, the USM slider will
start to slip, and occasionally causes audible sound noise.
Suppose the elliptic locus of a point on the stator surface is expressed by
u = u0 sin(ω 0 t + α ), and w = w 0 sin(ω 0 t + β), (P7.11.2)
the vertical speed at the top of this elliptic locus is given by
(∂u /∂t ) = ω 0u0 cos(ω 0 t + α ). (P7.11.3)
The maximum slider speed after reaching the steady motion is close to ω0 u 0. General remarks
for the speed: (1) if a higher order harmonic mode such as 3ω0 is used, higher speed is measured.
However, since using a higher order harmonic usually decreases the displacement u, too much
higher order harmonic modes again decrease the speed. (2) when the pressing force N is small,
the maximum displacement u is obtained, leading to the maximum speed. With increasing N, the
displacement u and the speed are reduced, contrary to the thrust increase.
7.8.3 STANDING WAVE–TYPE MOTORS

7.8.3.1 Rotary Motors
T. Sashida developed a rotary-type motor similar to the fundamental structure in Figure 7.52.71 Four
vibratory pieces were installed on the edge face of a cylindrical vibrator, and pressed onto the rotor.
This is one of the prototypes which triggered the present active development of USMs. A rotation
speed of 1500 rpm, torque of 0.08 N.m, and an output of 12 W (efficiency 40%) were obtained under
an input of 30 W at 35 kHz. This type of USM can provide a speed much higher than the inch-
worm types, because of its high operating frequency and amplified vibration displacement at the
resonance frequency.
Hitachi Maxel significantly improved the torque and efficiency by using a torsion coupler replac-
ing Sashida’s vibratory pieces (Figure 7.54), and by increasing the pressing force with a bolt.72 The
torsion coupler looks like an old-fashioned TV channel knob, consisting of two legs which trans-
form longitudinal vibration generated by the Langevin vibrator to a bending mode of the knob disk,
and a vibratory extruder. Notice that this extruder is aligned with a certain cant angle to the legs,
which transforms the bending to a torsion vibration. This transverse moment coupled with the bend-
ing up-down motion leads to an elliptical rotation on the tip portion, as illustrated in Figure 7.54b.
The optimum pressing force to get the maximum thrust is obtained, when the ellipse locus is
deformed roughly by half. A motor 30 × 60 mm2 in size and with a 20° ~ 30° in cant angle between
leg and vibratory piece can generate torques as high as 1.3 N.m with an efficiency of 80%. However,
this type provides only unidirectional rotation. Note also that even though the drive of the motor is
intermittent, the output rotation becomes very smooth because of the inertia of the rotor.
In collaboration with Samsung Electromechanics, Korea, The Penn State University developed
a zoom mechanism with two microrotary motors.73–75 A micromotor called “metal tube type” con-
sisting of a metal hollow cylinder and two PZT rectangular plates was used as basic microactuators
(See Figure 7.55a). When we drive one of the PZT plates, Plate X, a bending vibration is excited
basically along x′-axis. However, because of an asymmetrical mass (Plate Y), another hybridized

Propeller
Torsion coupler
A1 horn
Spacer
Piezoelectric
disk
Al cylinder
(a) (b)
FIGURE 7.54 Torsion coupler USM (a) and the motion of the torsion coupler (b).
Y
Elastic hollow
cylinder
X' Plate X
Y'
Plate Y
(a)
FIGURE 7.55 “Metal tube motor” using a metal tube and two rectangular PZT plates. (a) Schematic struc-
ture and (b) photo of the world’s smallest motor (1.5 mmφ).
bending mode is excited with some phase lag along y'-axis, leading to an elliptical locus in a clock-
wise direction, like a Hula-Hoop motion. In order to obtain a counterclockwise rotation, Plate Y is
now driven. The rotor of this motor is a cylindrical rod with a pair of stainless ferrule pressed with
a spring. The assembly is shown in Figure 7.55b. The metal cylinder motor 2.4 mm in diameter and
12 mm in length was driven at 62.1 kHz in both rotation directions. A no-load speed of 1800 rpm and
an output torque up to 1.8 N m were obtained for rotation in both directions under an applied rms
voltage of 80 V. Rather high maximum efficiency of about 28% for this small motor is a noteworthy
feature.
The world’s smallest camera with both optical zooming and autofocusing mechanisms for a cel-
lular phone application developed in 2003 is shown in Figure 7.56.75 Two micro-USMs with 2.4 mm
diameter and 14 mm length were installed to control zooming and focusing lenses independently in
conjunction with screw mechanisms.
NEC-Tokin developed a piezoelectric ceramic cylinder for a torsion vibrator (Figure 7.57).76 Using
an interdigital-type electrode pattern printed with a 45° cant angle on the cylinder surface, torsion
vibration was generated, which is applicable for a simple USM (refer to Example Problem 2.3).

FIGURE 7.56 Camera auto zooming/focusing mechanism with two metal tube USMs in a Samsung cellular
phone.
Ceramic
hollow-cylinder
t 45°
p
a
(a) (b)
FIGURE 7.57 Piezoelectric cylinder torsion vibrator (a) and its electrode pattern (b).
Shaft
Ball
bearing Spring
Rotor
Stator
head
Longitudinal
PZT
Torsional
PZT
Bottom
NUT
Bolt
FIGURE 7.58 Two-vibration-mode coupled-type motor.
Ueha proposed a two-vibration-mode coupled type motor (Figure 7.58). In this design, a tor-
sional Langevin vibrator was combined with three ML actuators to generate larger longitudinal
and transverse surface displacements of the stator, as well as to control their phase difference.77
The phase change can change the rotation direction.
7.8.3.2 Linear Motors

Uchino invented a π-shaped linear motor.78 This linear motor is equipped with a ML piezoelectric
actuator and fork-shaped metallic legs as shown in Figure 7.59. Since there is a slight difference in
the mechanical resonance frequency between the two legs, the phase difference between the bending
vibrations of both legs can be controlled by changing the drive frequency. The walking slider moves
in a way similar to a horse using its fore and hind legs when trotting. A test motor, 20 × 20 × 5 mm3

To oscillator 90˚
0T
Multilayer
piezoelectric
Adhesion actuator 1/4 T
Leg 2/4 T
Rail
3/4 T
1T
(a) (b)
FIGURE 7.59 π-shaped linear USM. (a) Construction and (b) walking principle. Note the 90° phase differ-
ence of two legs similar to that associated with human walking.
v
p
15
7
Velocity v (cm/s)
6
Efficiency p (%)
10 5
4
3
5
2
1
m m
0 50 100 0 50 100
Load m(gf) Load m(gf)
FIGURE 7.60 Motor characteristics of the π-shaped motor.
in dimension, exhibits a maximum speed of 20 cm/s and a maximum thrust of 0.2 kgf (=2 N) with a
maximum efficiency of 20%, when driven at 98 kHz at 6 V (actual power = 0.7 W). Figure 7.60 shows
the characteristics of the linear motor. This motor has been employed in a precision X–Y stage.
7.8.4 PROPAGATING WAVE–TYPE MOTORS

7.8.4.1 Linear Motors
Sashida and Ueha et al. manufactured a linear motor as illustrated in Figure 7.61.79,80 Two piezo-
electric vibrators installed at both ends of a steel transmission rod excite and receive the traveling
transverse wave (antisymmetric fundamental Lamb wave mode). Adjusting a load resistance in the
receiving vibrator leads to a perfect traveling wave. Exchanging the roles of the transmitting and
receiving piezocomponents results in a reversal of the movement.
Using the bending vibration, the wavelength λ can be chosen as short as several millimeters to
satisfy a stable surface contact with the slider by changing the cross-section area or the moment of
inertia of the transmission rod. In the case of Figure 7.61, λ = 26.8 mm. A slider, the contact face
of which is coated with rubber or vinyl resin, clamps the transmission rod with an appropriate
force. The transmission efficiency is strongly affected by the vibration source position on the rod,
and shows a periodic variation with the distance from the free end of the rod to the position of the

3 × 6 mm
Transmission rod l
6 cm
Horn (1:4) Slider
Piezovibrator
28 kHz
20 ø
L L
~
R
FIGURE 7.61 Linear motor using a bending vibration.
vibrator. Taking into account the wave phase, the vibration source should be fixed at a distance
corresponding to one wavelength λ (i.e., 26.8 mm) from the rod end.
The slider, made of a steel clamper 60 mm in length, which theoretically covers two waves, was
driven at a speed of 20 cm/s with a thrust of 50 N at 28 kHz. A major problem with this type of motor
is found in its low efficiency (around 3%), because the whole rod must be excited even when only a
small portion is utilized for the output. To overcome this dilemma, ring-type motors were invented,
where the whole rod can be utilized, because the lengths of the stator and rotor are the same.
7.8.4.2 Rotary Motors

When we deform the rod discussed in the previous section to make a ring by connecting the two
ends topologically, we can make a rotary-type motor using a bending vibration. Two types of “ring”
motor designs are possible: (a) the bending mode type and (b) the extensional mode type.81 Although
the principle is similar to the linear type, more sophisticated structures are employed with respect
to the ceramic poling and the mechanical support mechanism.
In general, when a vibration source drives one position of a closed ring (circular or square) at a
frequency corresponding to the resonance of this ring, only a standing wave is excited, because the
vibration propagates in two directions symmetrically from the vibration source and interference
occurs. When multiple vibration sources are installed on the ring, displacements can be obtained
by superimposing all the waves (two waves from each vibration source). Using the superimposition
principle, we can generate a traveling-like wave in the closed ring with the profile of the original
standing wave.
Refer to Figure 7.62. Assuming a vibration source of A cos ωt at the point θ = 0 of the elastic ring,
the nth mode standing wave can be expressed by
u(θ, t ) = A cos nθ cos ωt , (7.50)
and the traveling wave by
u(θ, t ) = A cos(nθ − ωt ). (7.51)
Since the traveling wave can be expressed as a superimposition of two standing waves as
u(θ, t ) = A cos nθ cos ωt + A cos(nθ − π /2) cos(ωt − π /2), (7.52)
we can derive an important principle: a propagating wave can be generated by superimposing two
standing waves whose phases differ by 90° to each other both in time and in space. More generally,
the phase difference can be chosen arbitrarily (except 0, −π, π), as long as the phase difference is
maintained in space and time. The vibration source positions for generating a propagating wave on

cos ωt
t
sω ωt
co cos
sin ωt
sin ωt sin ωt
u u u
θ θ θ
sin cos sin cos sin cos
(a) 1st (b) 2nd (c) 3rd
FIGURE 7.62 Vibration source positions for generating a propagating wave in a ring.
a ring are illustrated in Figure 7.62. In principle, excitation at only two parts of the ring is sufficient
to generate a traveling wave. However, in practice, the number of vibration sources is increased to
as many as possible to increase the mechanical output. The symmetry of the electrode structure also
needs to be considered.
Figure 7.63 shows the famous Sashida motor.82 By means of the traveling elastic wave (up-down
motion) induced by a thin piezoelectric ring (i.e., unimorph type), a ring-type slider in contact with
the “rippled” surface of the elastic body bonded onto the piezoelectric is driven in both directions by
exchanging the sine and cosine voltage inputs. Another advantage is its thin design, which makes it
suitable for installation in cameras as an automatic focusing device. Eighty percent of the exchange
lenses in Canon’s “EOS” camera series have already been replaced by the USM mechanism.
The PZT piezoelectric ring is divided into 16 positively and negatively poled regions and two
asymmetric electrode gap regions so as to generate a 9th-mode propagating wave at 44 kHz. [Two
comments: (1) In order to escape from the audible frequency range, a large size motor should be
operated at a high-order harmonic mode. (2) Since too high-order harmonics reduce the speed and
torque, choose a reasonable mode]. A prototype was composed of a brass ring of 60 mm in outer
diameter, 45 mm in inner diameter, and 2.5 mm in thickness, bonded onto a PZT ceramic ring of
0.5 mm in thickness with divided electrodes on the backside. The rotor was made of polymer coated
with hard rubber or polyurethane. Figure 7.64 shows Sashida’s motor characteristics.82
Canon utilized the Sashida’s “surfing” motor for a camera automatic focusing mechanism,
installing the ring motor compactly in the lens frame. It is noteworthy that the stator elastic ring has
many teeth, which can magnify the transverse elliptical displacement and improve the speed. The
lens position can be shifted back and forth with a screw mechanism. The advantages of this motor
over the conventional electromagnetic motor are
1. Silent drive due to the ultrasonic frequency drive and no gear mechanism (i.e., more suit-
able for video cameras with microphones)
2. Thin motor design and no speed reduction mechanism such as gears, leading to space saving
3. Energy saving

Rubber Brass ring

PZT E
Rotor
+
– + Polarization
–
–
(3/4)λ +
φ 45
φ 60
+ –
– + B
A + (1/4)λ –
– + V0 sin ωt
V0 cos ωt + –
3 0.5 λ
0.5 2.5
6.5
(a)
Axis
Rotor
Slider
Elastic ring
Piezoring
Bearing Felt
FIGURE 7.63 Stator structure of Sashida’s motor. (From Sashida, T., Mech. Automat. Jpn., 15, 31, 1983.)
1 gf = 9.8×10–3N A
40 V0 sin ωt
B 1.5 mH
V0 cos ωt
Applied voltage
30
V E
Revolution (rpm)
0 =1
0.3 Ground
V
V
0 =6 f = 44 khz
20 .5V
V
0 =4
.6V
10
0 200 400 600 800 1,000 1,200

Torque (gf cm)
FIGURE 7.64 Motor characteristics of Sashida’s motor.
A general problem encountered for these traveling wave–type motors is the support mechanism of
the stator. In the case of a standing wave motor, the nodal points or lines are generally supported;
this causes minimum effects on the resonance vibration. A traveling wave, however, does not have
such steady nodal points or lines. Thus, special considerations are necessary. In Figure 7.63, the

Spring for
Rotor pressure
Stator
vibrator
Lead Support plate Piezoelectric

wire for stator ceramic
(a) (b)
FIGURE 7.65 (a) Construction of Seiko’s motor and (b) a photo of 10 mm and 4 mmφ motors.
stator is basically supported very gently along the axial direction or felt so as not to suppress the
bending vibration. It is important to note that the stop-pins which latch onto the stator teeth only
provide high rigidity against the rotation. Matsushita Electric proposed a nodal line support method
using a higher order vibration mode, where a wide ring is supported at the nodal circular line and
“teeth” are arranged on the maximum amplitude circle to get larger revolution.83
Seiko Instruments miniaturized the USM to dimensions of 10 mm and 4 mmφ in diameter using
basically the same principle.84 Figure 7.65a shows the construction of one of these small motors with
a 10 mm diameter and a 4.5 mm thickness. A driving voltage of 3 V and a current of 60 mA pro-
duces 6000 rev/min (no-load) with a torque of 0.1 mN.m. A 10 mm motor in Figure 7.65b has been
installed in a wristwatch for silent alarm function, while a 4 mm motor is used for perpetual func-
tion (i.e., date plate change). AlliedSignal developed USMs similar to Shinsei’s, which are utilized
as mechanical switches for launching missiles.85
It is important to note that the unimorph (a piezoceramic plate and a metal plate bonded together)
bending actuation cannot provide high efficiency theoretically, because the electromechanical cou-
pling factor k is usually less than 10%. Therefore, instead of the unimorph structure, a simple disk
was directly used to make motors.86,87 Figure 7.66 shows (1, 1), (2, 1) and (3, 1) modes of a simple
disk, which are axial-asymmetric modes. Both the inner and outer circumferences can provide a
rotation like a “hula hoop.”
7.8.4.3 Comparison of the Various Ultrasonic Motors

The standing-wave type, in general, is low in cost (one vibration source) and has high efficiency (up
to 98% theoretically), but lack of control in both the clockwise and counterclockwise directions is
a drawback. In comparison, the propagating-wave type combines two standing waves with a 90°
phase difference both in time and space. This type requires, in general, two vibration sources to
(a) ((1, 1)) mode (b) ((2, 1)) mode (c) ((2, 2)) mode (d) ((3, 1)) mode
sin ωt sin ωt sin ωt

cos ωt
cos ωt
cos ωt
(a´) ((1, 1)) mode (b´) ((2, 1)) mode (d´) ((3, 1)) mode
(c´) ((2, 2)) mode
FIGURE 7.66 Disk-type hula-hoop motor.

Drive friction rod ф 1.0 × 6.0 mm

Multilayer piezo ф 1.2 × 2.3 mm
Fixture ф 2.5 × 0.6 mm
Fixture
Multilayer piezo
Drive friction rod

Size: ф2.5×8.9 mm
FIGURE 7.67 Illustration of the SIDM developed by Konica-Minolta for phone camera zooming applica-
tions. (From Okamoto, Y. et al., Konica Minolta Tech. Rep., 1, 23, 2004. With permission.)
Piezo
Displacement
Slider
Time
FIGURE 7.68 Displacements of the piezoelement and the slider to show the “stick & slick” motion. (From
Okamoto, Y. et al., Konica Minolta Tech. Rep., 1, 23, 2004. With permission.)
generate one propagating wave, leading to low efficiency (not more than 50%), but it is controllable
in both rotational directions.
7.8.5 SMOOTH IMPACT DRIVE MECHANISM (SIDM)

Though the principle is different from the USM, a competitive technology is an impulse motor.
Konika-Minolta developed a smooth impact drive mechanism (SIDM) using a ML piezoelement.88
The idea comes from the “stick & slick” condition of the ring object attached on a drive rod in
Figure 7.67. By applying a saw shape voltage to a ML actuator, alternating slow expansion and quick
shrinkage are excited on a drive friction rod (see Figure 7.68). A ring slider placed on the drive
rod will stick on the rod due to friction during a slow expansion period, while it will slide during
a quick shrinkage period, so that the slider moves from the bottom to the top. A lens is attached to
this slider. In order to obtain the opposite motion, the voltage saw shape is reversed. Compared to
the USMs, the impulse motor is simpler, but the 1/10 smaller holding force may be a problem. The
drive frequency can be much higher than the inchworm type, close up to the resonance frequency
(i.e., the frequency inaudible to the human).
7.9 PIEZOELECTRIC ENERGY HARVESTING

7.9.1 PIEZOELECTRIC PASSIVE DAMPING TO ENERGY HARVESTING
As is discussed in Section 10.4, piezoelectric dampers have been developed by the author’s group
in 1980s. In order to suppress mechanical noise vibration, Uchino et al. used a piezoelectric, which
could convert the vibration energy to electric energy. The resistive shunt method was patented
for consuming the converted electric energy in Joule heat, so that the mechanical vibration was
dramatically damped. A piezoceramic patch was bonded on an elastic plate with its fundamental

Short 13 KΩ
1 KΩ 54 KΩ
40
Displacement (μm)
20
6.6 KΩ Open
0
0 40 80 120 160 0 40 80 120 160
Time (ms) Time (ms)
FIGURE 7.69 Resistive shunt piezoelectric patch for vibration damping test. The fundamental resonance
frequency of this vibration system is around 200 Hz. Note that the resistance 6.6 kΩ exhibits the quickest
damping.
resonance frequency around 200 Hz. Figure 7.69 shows the fact that there is an optimal resistance
6.6 kΩ to be connected to the piezoelectric damping device in order to obtain the quickest damping;
that is, the resistance should be chosen equal to the piezodamper impedance Z = 1/ωC (impedance
matching).89 The piezo-shunt methodologies have now been extended to a combination of resistive,
capacitive, and inductive components.
Consuming the converted electric energy in Joule heat seemed to be useless. Thus, Uchino
et al. started to accumulate the electric energy in batteries in the 1990s. Unused power exists in vari-
ous forms such as vibrations, water flow, wind, human motion, and shock waves. In recent years,
industrial and academic research units have focused attention on harvesting energy from vibrations
using piezoelectric transducers. These efforts have provided the initial research guidelines and have
brought in light the problems and limitations of implementing the piezoelectric transducer. There
are three major steps associated with piezoelectric energy harvesting90:
i. Mechanical–mechanical energy transfer

This includes mechanical stability of the piezoelectric transducer under large stresses, and
mechanical impedance matching.
ii. Mechanical–electrical energy transduction
This relates with the electromechanical coupling factor in the composite transducer
structure.
iii. Electrical–electrical energy transfer
This includes electrical impedance matching, a DC/DC converter to accumulate the energy
into a capacitor or a rechargeable battery.
7.9.2 HIGH-ENERGY HARVESTING (~W)

Periodic vibrations generated from rotating machines or engines are one of the most promising
sources for recovering energy. Kim et al. investigated first the capability of harvesting the electrical
energy from mechanical vibrations in dynamic environment, such as an automobile engine through

Engine
Bolt to support
W
T the engine
Change output
PZT L
FIGURE 7.70 High power energy harvesting (1 W at 100 Hz) with “Cymbal” piezoelectric flexible
composites.
a “Cymbal” piezoelectric transducer.91–93 The targeted mechanical vibration from an engine lies
in the range of 50 ~ 150 Hz with force amplitude of the order of 1 kN. It was found that under such
severe stress conditions the metal–ceramic composite transducer “Cymbal” is a promising rigid
structure keeping a relatively high electromechanical transduction rate. The metal cap enhances
the endurance of the ceramic to sustain high loads along with stress amplification. In our study, the
experiments were performed under force of 7 ~ 70 N at the frequency of 100 Hz on a cymbal with
29 mm diameter and 1 mm thickness. At this frequency and force level, 60 mW power was generated
from a cymbal measured across a 400 kΩ resistor. Figure 7.70 illustrates the final cymbal energy
harvesting composites, in which three cymbals are embedded in a triangular shape rubber sheet.
These composite sheets are inserted as mats below the engine. Since each cymbal can generates
close to 100 mW continuously, 1 W can be obtained in total. Note that the optimal conditions for
obtaining the maximum vibration damping and the maximum energy harvesting are exactly the
same; that is, we can chase two “rabbits” at the same time.
Usually the output impedance of the harvested electrical energy with a piezoelectric transducer
is extremely high, so that the electrical impedance mismatch provides significant reduction of effi-
ciency for accumulating the electric charge into a rechargeable battery, which has typically low
impedance around 10 ~ 100 Ω. To overcome this problem, so-called switch-shunt is most popularly
used for the energy harvesting purpose. A DC–DC buck converter in Figure 7.71 was designed for
the above cymbal system which allows transfer of 50 mW power out of 60 mW from the Cymbal
(82% efficiency) to a low impedance load of 5 kΩ with a 2% duty cycle and at a switching frequency
of 1 kHz.93
IR2106
Gate drive
L
RG154 33 mH
R load
C1 10 μF C2
(rechargeable
100 μF
IN914 battery)
Cymbal transducer
FIGURE 7.71 Piezoelectric energy harvesting circuit with a switch shunt, DC–DC buck converter.

Charging the
electronic
Flexible
piezoelectric
textile
Flexible energy
harvest circuit
FIGURE 7.72 IC energy harvesting system with MFC.
Another unique circuit design may be with a piezoelectric transformer. The piezoelectric trans-
former used in the circuit has low output impedance around 50 Ω, and the efficiency of the piezo-
electric transformer in the resonance shows over than 95%. This low output impedance is suitable
for an impedance matching to the load (rechargeable battery).94
7.9.3 LOW-ENERGY HARVESTING (~MW)

Another promising vibration source is a human motion. Uchino et al. are developing intelligent
clothing (IC) with piezoelectric energy harvesting system of flexible piezoelectric textiles, aiming at
a general power source for charging up portable equipment such as cellular phones, health monitor-
ing units, or medical drug delivery devices. The macro fiber composite (MFC) is an actuator that
offers high performance and flexibility in a cost-competitive device (Smart Material Corp.). The
MFC consists of rectangular piezoceramic rods sandwiched between layers of adhesive and elec-
troded polyimide film. This film contains interdigitated electrodes that transfer the applied voltage
directly to and from the ribbon-shaped rods. This assembly enables in-plane poling, actuation, and
sensing in a sealed, durable, ready-to-use package. When embedded in a surface or attached to flex-
ible structures, the MFC actuator provides distributed solid-state deflection and vibration control
(see Figure 7.72). The MFC composites can generate power in milliwatts.90
Remote electric switch developed by Face/PulseSwitch Systems, LC, VA is one of the successful
products using a unimorph-type (Thunder) piezodevice.95 Lightning switch wireless transmitters
(the switch controls) use NASA space technology to generate their own electricity whenever the
transmitter button is pushed mechanically through the piezoelectricity. Their radio signals go 40′ to
100+′ right through walls, floors and ceilings.

Let us consider the energy flow/conversion from the input mechanical energy source to the out-
put rechargeable battery, based on Figures 7.70 and 7.71. The sample was a cymbal with 0.3 mm
thick stainless steel endcaps, inserted below a 4 kg engine weight (40 N bias force). The engine
shook at 100 Hz for 8 s, and the accumulated energy during that time period was measured at
each point. The energy was summarized in Table 7.6. Discuss why the energy amount decreases
with transmitting and transducing successively with numerical analyses.
Solution
1. Source to transducer: Mechanical impedance mismatch

8.22 J/9.48 J = 87% of the mechanical energy is transmitted from the source to piezocym-
bal transducer. If the mechanical impedance is not seriously considered in the transducer

TABLE 7.6
Energy Flow/Conversion Analysis in the Cymbal Energy Harvesting Process
Source Æ Transducer Æ Transducer Æ Circuit-in Æ Battery
Mechanical Mechanical Electrical Electrical Electrical
Energy ¨ Energy ¨ Energy ¨ Energy ¨ Energy
9.48 J 8.22 J 0.74 J 0.42 J 0.34 J
(too mechanical-soft or hard), the mechanical energy will not transfer efficiently, and
large amount will be reflected at the contact point.
2. Transduction in the transducer: Electromechanical coupling
Transduction rate is evaluated by k2. Since the keff = 0.25 ~ 0.30 in the cymbal, the energy
conversion rate will be 9% maximum. The remaining portion will remain as the original
mechanical vibration energy. 0.74 J/8.22 J = 9.0%.
3. Transducer to harvesting circuit: Electrical impedance matching
0.42 J/0.74 J = 57% is related to the electrical impedance mismatch between the cymbal
output and the circuit input.
4. Harvesting circuit to rechargeable battery:
0.34 J/0.42 J = 81%. This reduction partially originates from the energy consumption in the
circuit, and partially from the electrical impedance mismatch between the circuit output
(still around 5 kΩ) and the rechargeable battery impedance (around 10–20 Ω).
In summary, 0.34 J/9.48 J = 3.6% is the current energy harvesting rate in the battery in com-
parison with the initial source vibration energy. The key to dramatic enhancement is to use a
high-k mode; that is, k33, kt, or k15.
Chapter Essentials
1. Piezoelectric figures of merit:

a. Piezoelectric strain constant d—x = dE
b. Piezoelectric voltage constant g—E = gX
c. Electromechanical coupling factor k—k2 = (stored mechanical energy/input electrical
energy) = d2/ε0εs
d. Mechanical quality factor—Q M = ω0/2Δω
e. Acoustic impedance—Z = (ρc)1/2
2. Piezoelectric equations:
xi = sijE X j + dmi Em ,
Dm = dmi Xi + εXmk Ek (i, j = 1,2,...,6; m, k = 1,2,3).
3. The resonance and antiresonance modes are both mechanical resonances. Electrical admit-
tance maximum and minimum correspond to resonance and antiresonance, respectively.
The resonance frequencies are ω0, 3ω0, 5ω0, … (odd number multiplication of the funda-
mental mode frequency).

4. Classification of ceramic actuators:
Displacement Drive Technique Actuator Category Material

Rigid displacement Servo drive Servo displacement transducer Electrostrictor
On/off drive Pulse-drive motor Soft piezoelectric
Resonant displacement AC resonance drive USM Hard piezoelectric
5. Merits and demerits of the USMs:

Merits
1. Low speed and high torque—Direct drive
2. Quick response, wide velocity range, hard brake and no backlash
–Excellent controllability
–Fine position resolution
3. High power/weight ratio and high efficiency
4. Quiet drive (driven at an inaudible frequency for human)
5. Compact size and lightweight
6. Simple structure and easy production process
7. Negligible effect from external magnetic or radioactive fields, and also no generation
of these fields
Demerits
8. Necessity for a high-frequency power supply
9. Less durability due to frictional drive
10. Drooping torque vs. speed characteristics
6. How to generate a traveling wave on an elastic ring:
nth mode standing wave: u(θ, t) = A cos nθ ⋅ cos ωt
nth mode travelling wave: u(θ, t ) = A cos(nθ − ωt )
= A cos nθ ⋅ cos ωt + A cos(nθ − π / 2) ⋅ cos(ωt − π / 2).
A propagating wave can be generated by superimposing two standing waves whose phases
differ by 90° to each other both in time and in space.
7. Three major steps associated with piezoelectric energy harvesting:
i. Mechanical–mechanical energy transfer
This includes mechanical stability of the piezoelectric transducer under large stresses and
mechanical impedance matching.
ii. Mechanical–electrical energy transduction
This relates to the electromechanical coupling factor in the composite transducer structure.
iii. Electrical–electrical energy transfer
This includes electrical impedance matching, a DC/DC converter to accumulate the energy
into a capacitor or a rechargeable battery.
Check Point
1. (T/F) PMN electrostrictive ceramic is the most suitable actuator material for an ink-jet
printer head application. True or False?
2. (T/F) When 1 W electric energy is input to a piezoelectric material with an electrome-
chanical coupling factor k = 0.71 (71%), we can expect 0.5 W mechanical energy converted

in this piezoelectric material. Thus, supposing that no loss is considered in this actuator
system, we can conclude that the efficiency of this device is 50%. True or False?
3. (T/F) When a piezoelectric actuator is driven by a rectangular pulse voltage, the mechani-
cal ringing is completely suppressed when the pulse width is adjusted exactly to half of the
resonance period of the sample. True or False?
4. (T/F) The resonance point of a piezoelectric sample can be identified by an impedance
maximum peak of the V vs. I relation. True or False?
5. Fill in the blank (a) in the following equation, which is a key to designing the USM:
cos kx cos ωt + cos(kx − π /2) cos(ωt − π /2) = cos[(a)].
Chapter Problems
7.1 Calculate the electromechanical coupling factor kij of a piezoelectric ceramic vibrator
for the following vibration mode:
a. Length extension mode through d31 (⊥ E): d31
b. Shear mode on the plate: k15
Hint
Necessary equations are
(a) x1 = s11
E
X1 + d31E3,
D3 = d31 X1 + ε33
X
E3 .
(b) x5 = s55
E
X 5 + d15 E1,
D1 = d15 X 5 + ε11
X
E1.
7.2 Using a rectangular piezoceramic plate (length: L; width: w; and thickness: b; poled
along the thickness. Figure 7.14), determine the electromechanical coupling parameters
k31, d31, and Q M. Explain the processes for both the resonance and pulse-drive methods.
The density ρ and permittivity ε33 X of the ceramic must be known prior to the following
experiments.
a. Using an impedance analyzer, the admittance for the mechanically free sample
(supported at the nodal point, i.e., at the center of the plate) was measured as a func-
tion of the drive frequency ω and the admittance curve was obtained. Explain how
to determine the k31, d31, and Q M values from the experimental data. Also verify that
the following approximate equation can be used for a low-coupling piezoelectric
material (refer to Equation 7.33):
2
k31/(1 − k31
2
) = (π 2/4)( Δf /fR ) ( Δf = fA − fR ).
b Using a pulse-drive technique, the transient displacement was measured as a func-

tion of time, and the displacement curve was obtained (See Figure 7.73). Explain an
alternative way how to determine the k31, d31, and Q M values from the data.
Hint
Refer to Section 7.3.1 for the first part. The pulse-drive method is an alternative method
for measuring high voltage piezoelectric characteristics. By applying a step electric field
to a piezoelectric sample, the transient vibration corresponding to the desired mode
(extensional, bending, etc.) is measured. The resonance period, stabilized displacement

Applied voltage
(V)
Displacement
(μm)
Time
Resonance period
FIGURE 7.73 Pulse-drive technique for measuring the electromechanical parameters.
and damping constant are obtained experimentally, from which the elastic compliance,
piezoelectric constant, mechanical quality factor [use the relation: Q M = (1/2)ω0τ] and elec-
tromechanical coupling factor can be calculated. Although the experimental accuracy is
not high, the simple setup is attractive especially for its low cost. Moreover, unlike the
resonance/antiresonance methods, this technique requires only one voltage pulse, and thus
does not generate heat. Thus, the electric field dependence of piezoelectricity can be mea-
sured. The background theory of this technique is in Example Problem 7.10.
7.3 Using a rectangular piezoelectric plate and the transverse d31 mode, design a flight actu-
ator (pinball machine).
a. When a rectangular negative pulse (−E0) is applied to a plate sample installed nor-
mally and fixed rigidly at one end, verify that the velocity at the other end is given
by 2|d31|E 0v (v: sound velocity in the ceramic), and is independent of the length.
b. Suppose that this velocity is acquired through a small steel ball (mass: M) without
loss. Calculate the maximum height of the steel ball, when the ball is hit exactly
vertically.
7.4 From the strain distribution x1(x) for a low electromechanical coupling material pictured
in Figure 7.14, draw the displacement distribution u(x) for both the resonance and anti-
resonance states, and discuss the difference between the two states.
7.5 For the equivalent circuit of a piezoelectric transducer at the antiresonance state (Figure
7.17b), derive the relations of L and C to the intrinsic physical parameters such as ρ, d,
sE, and the dimensions of the transducer.
7.6 A piezotransformer (2nd resonance mode usage) is held on a circuit board as shown in
the photo below (Figure 7.74a). Two ring materials are surrounded on the transformer
and the bottom of these two rings are adhered on the board. Consider the transformer
installed in the following sequence:
a. What kind of material should be chosen for the rings? Why?
b. Where should the rings be placed? Why?
c. What kind of lead wire (metal species, size, etc.) should be used? Why (from both
electrical and mechanical viewpoints)?
d. The lead wires in this backlight inverter product are connected on both edges of the
PZT rectangular plate (White-resin-drop portions). Is the present lead wire position
recommended? Otherwise, where should the lead wire be installed? Why?
For your reference, Figure 7.74b shows the configuration of this piezoelectric transformer
and the second resonance mode.
Hint
A piezotransformer (2nd resonance mode usage) has two nodal lines; one is ¼ of the full
length from the left edge, the other ¼ from the right edge, as shown in Figure 7.74a. The
nodal point does not displace, but the maximum strain and stress are concentrated.

FIGURE 7.74 (a) Configuration of a backlight inverter with a modified Rosen-type piezoelectric transformer,
and (b) polarization direction and displacement distribution contour for the 2nd resonance mode.
a. Two rings should be made of a soft material, because the mechanical mismatch
with PZT is essential not to dissipate the vibration energy. What will be the material
candidates?
b. The rings should be placed around the nodal lines, because clamping at the nodal
point shows the minimum disturbance of the resonating vibration, not to suppress
the vibration displacement.
c. Electrical viewpoint
Because the electric impedance of the transformer at the resonance point is very
small, the resistance of the lead wire should also be minimized. The wire material
should have high conductivity. What will be practical wire candidates?
Mechanical viewpoint
Because the mass of the lead wire affects also drastically the vibration mode, we
should minimize the wire thickness. However, too thin wire decreases the conduc-
tance, we need to compromise the thickness.
d. Similar to (b), the lead wires should be placed around the nodal lines, because
mass/load at the nodal point shows the minimum disturbance of the resonating
vibration, not to suppress the vibration displacement. From this reason, is the pres-
ent position recommended?
7.7 Describe the general principle for generating a traveling wave on an elastic ring using a
unimorph structure, like in Figure 7.63. Then, taking the second-order vibration mode
u(θ, t) = A cos(2θ − ωt), describe practical poling and electrode configurations to be
applied to the piezoceramic ring for generating this mode of vibration. (There will be
at least two different configurations.)
REFERENCES
1. B. Jaffe, W. Cook, and H. Jaffe: Piezoelectric Ceramics, Academic Press, London (1971).
2. W. G. Cady: Piezoelectricity, McGraw-Hill, New York, revised edition by Dover Publications (1964).
3. M. E. Lines and A. M. Glass: Principles and Applications of Ferroelectric Materials, Clarendon Press,
Oxford (1977).
4. K. Uchino: Piezoelectric Actuators and Ultrasonic Motors, Kluwer Academic Publishers, Boston, MA
(1996).
5. K. Uchino: Piezoelectric/Electrostrictive Actuators, Morikita Publishing, Tokyo (1986).

6. T. Ikeda: Fundamentals of Piezoelectric Materials Science, Ohm Publishing Co., Tokyo (1984).
7. Y. Ito and K. Uchino: Piezoelectricity, in Wiley Encyclopedia of Electrical and Electronics Engineering,
Vol. 16, p. 479, John Wiley & Sons, New York (1999).
8. W. A. Smith: Proc. SPIE—The Int’l Soc. Opt. Eng., 1733 (1992).
9. H. Takeuchi, S. Jyomura, E. Yamamoto, and Y. Ito: J. Acoust. Soc. Am., 74, 1114 (1982).
10. Y. Yamashita, K. Yokoyama, H. Honda, and T. Takahashi: Jpn. J. Appl. Phys., 20(Suppl. 20-4), 183
(1981).
11. Y. Ito, H. Takeuchi, S. Jyomura, K. Nagatsuma, and S. Ashida: Appl. Phys. Lett., 35, 595 (1979).
12. H. Takeuchi, H. Masuzawa, C. Nakaya, and Y. Ito: Proc. IEEE 1990 Ultrasonics Symp., 697 (1990).
14. T. R. Shrout, Z. P. Chang, N. Kim, and S. Markgraf: Ferroelect. Lett., 12, 63 (1990).
15. R. E. Newnham, D. P. Skinner, and L. E. Cross: Mater. Res. Bull., 13, 525 (1978).
16. W. A. Smith: Proc. 1989 IEEE Ultrasonic Symp., 755 (1989).
17. Kistler, Stress Sensor, Production Catalog, Switzerland.
18. Tokin, Gyroscope, Production Catalog, Japan.
19. K. Uchino, S. Nomura, L. E. Cross, S. J. Jang, and R. E. Newham: Jpn. J. Appl. Phys., 20, L367 (1981);
K. Uchino: Proc. Study Committee Barium Titanate, XXXI-171-1067 (1983).
20. Y. Watabe: Bull. Ceramic Soc. Japan, 42(5), 396 (2007).
21. B. A. Auld: Acoustic Fields and Waves in Solids, 2nd ed., Robert E. Krieger, Melbourne (1990).
22. G. S. Kino: Acoustic Waves: Device Imaging and Analog Signal Processing, Prentice-Hall, Englewood
Cliffs, NJ (1987).
23. C. S. Desilets, J. D. Fraser, and G. S. Kino: IEEE Trans. Sonics Ultrason., SU-25, 115 (1978).
24. S. Saitoh, T. Takeuchi, T. Kobayashi, K. Harada, S. Shimanuki, and Y. Yamashita: Jpn. J. Appl. Phys.,
38(5B), 3380–3384 (1999).
25. http://www.shumurie.co.jp
26. Popular mechanics, 180(3), 20 (2003).
27. C. Campbell: Surface Acoustic Wave Devices and Their Signal Processing Applications, Academic Press,
San Diego, CA (1989).
28. H. Matthews: Surface Wave Filters, Wiley Interscience, New York (1977).
29. A. Yamamoto and T. Higuchi: Japan. Patent disclosure, P2004-190537A (2004).
30. Z.-Y. Cheng: Private communication, Auburn University, Auburn, AL (2003).
31. http://www.thinksrs.com/products/QCM200.htm
32. C. A. Rosen: Proc. Electronic Compon. Symp., p. 205 (1957).
33. S. Kawashima, O. Ohnishi, H. Hakamata, S. Tagami, A. Fukuoka, T. Inoue, and S. Hirose: Proc. IEEE
Int’l Ustrasonic Symp. ‘94, France (Nov. 1994).
34. K. Uchino: Proc. Actuator 2006, p. 48, Bremen Messe, Bremen, Germany (June 14–16, 2006).
35. K. Uchino: Bull. Am. Ceram. Soc., 65(4), 647 (1986).
36. K. Uchino: MRS Bull., 18(4), 42 (1993).
37. Handbook on New Actuators for Precision Position Control, Edit. in Chief K. Uchino., Fuji Technosystem,
Tokyo (1994).
38. K. Uchino: Recent developments in ceramic actuators, in Proc. Workshop Microsys. Technol. USA Can.,
Germany, mst news, special issue, VDI/VDE, 28–36 (1996).
39. K. Furuta and K. Uchino: Adv. Ceram. Mater., 1, 61 (1986).
40. J. von Cieminski and H. Beige: J. Phys. D, 24, 1182 (1991).
41. L. E. Cross, S. J. Jang, R. E. Newnham, S. Nomura, and K. Uchino: Ferroelectrics, 23(3), 187 (1980).
42. K. Uchino: Ceramic actuators, in Ceramic Data Book ‘88, Inst. Industrial Manufacturing Tech., Tokyo
(1988).
43. K. Uchino and S. Nomura: Ferroelectrics, 50(1), 191 (1983).
44. A. Furuta, K. Y. Oh, and K. Uchino: Sensors Mater., 3(4), 205 (1992).
45. S. Takahashi, A. Ochi, M. Yonezawa, T. Yano, T. Hamatsuki, and I. Fujui: Ferroelectrics, 50, 181
(1993).
46. A. Bauer and F. Moller: Proc. 4th Int’l Conf. New Actuators, Germany, p. 128 (1994).
47. Active Control Experts, Inc. Catalogue “PZT Quick Pack” (1996).
48. Y. Sugawara, K. Onitsuka, S. Yoshikawa, Q. C. Xu, R. E. Newnham, and K. Uchino: J. Am. Ceram. Soc.,
75(4), 996 (1992).
49. H. Goto, K. Imanaka, and K. Uchino: Ultrasonic Technol, 5, 48 (1992).
50. K. Uchino and S. Hirose: IEEE UFFC Trans, 48, 307–321 (2001).
51. N. Kanbe, M. Aoyagi, S. Hirose, and Y. Tomikawa: J. Acoust. Soc. Jpn. (E), 14(4), 235 (1993).

52. S. Sugiyama and K. Uchino: Proc. Int’l. Symp. Appl. Ferroelectr. ‘86, IEEE, p. 637 (1986).
53. C. Kusakabe, Y. Tomikawa, and T. Takano: IEEE Trans. UFFC, 37(6), 551 (1990).
54. T. Ota, T. Uchikawa, and T. Mizutani: Jpn. J. Appl. Phys., 24(Suppl. 24-3), 193 (1985).
55. J. T. Dorsey, T. R. Sutter, and K. C. Wu: Proc. 3rd Int’l Conf. Adaptive Struct., p. 352 (1992).
56. B. Wada: JPL Document D-10659, p. 23 (1993).
57. F. K. Straub: Smart Mater. Struct., 5, 1 (1996).
58. P. C. Chen and I. Chopra: Smart Mater. Struct., 5, 35 (1996).
59. A. Ohgoshi and S. Nishigaki: Ceramic Data Book ‘81, p. 35, Inst. Industrial Manufacturing Technology,
Tokyo (1981).
60. Okamoto et al.: Broadcast. Technol. (7), 144 (1982).
61. M. Kobayashi, M. Tomita, K. Yamada, and M. Matsumoto: Natl. Tech. Rep. 28, 419 (1982).
62. K. Yamada et al.: Abstract Television Soc. Jpn., 8–2 (1982).
63. T. Yano, E. Sato, I. Fukui, and S. Hori: Proc. Int’l Symp. Soc. Inf. Display, p. 180 (1989).
64. N. Kurashima: Proc. Machine Tech. Inst. Seminar, MITI, Tsukuba, Japan (1999).
65. Y. Yokoya: Electron. Ceram., 22(111), 55 (1991).
66. A. Fujii: Proc. JTTAS Meeting on Dec. 2, Tokyo (2005).
67. M. P. Koster: Proc. 4th Int’l Conf. New Actuators, Germany, p. 144 (1994).
68. H. V. Barth: IBM Techn. Discl. Bull., 16, 2263 (1973).
69. V. V. Lavrinenko, S. S. Vishnevski, and I. K. Kartashev: Izvestiya Vysshikh Uchebnykh Zavedenii,
Radioelektronica, 13, 57 (1976).
70. P. E. Vasiliev et al.: UK Patent Application GB 2020857 A (1979).
71. T. Sashida: Oyo Butsuri 51, 713 (1982).
72. A. Kumada: Jpn. J. Appl. Phys., 24(Suppl. 24-2), 739 (1985).
73. B. Koc, S. Cagatay, and K. Uchino: IEEE Ultrason. Ferroelect. Freq. Control Trans., 49(4), 495–500
(2002).
74. S. Cagatay, B. Koc, and K. Uchino: IEEE Trans. UFFC, 50(7), 782–786 (2003).
75. K. Uchino: Proc. New Actuator 2004 (Bremen, June14–16), p. 127 (2004).
76. Y. Fuda and T. Yoshida: Ferroelectrics, 160, 323 (1994).
77. K. Nakamura, M. Kurosawa, and S. Ueha: Proc. Jpn. Acoustic Soc., No. 1-1-18, p. 917 (Oct., 1993).
78. K. Uchino, K. Kato, and M. Tohda: Ferroelectrics, 87, 331 (1988).
79. S. Ueha and Y. Tomikawa: Ultrasonic motors, in Monographs in Electr. Electron. Eng., vol. 29, Oxford
Science Publ., Oxford (1993).
80. Nikkei Mechanical, Feb. 28 issue, p. 44 (1983).
81. M. Kurosawa, S. Ueha, and E. Mori: J. Acoust. Soc. Am., 77, 1431 (1985).
82. T. Sashida: Mech. Automat Jpn., 15(2), 31 (1983).
83. K. Ise: J. Acoust. Soc. Jpn., 43, 184 (1987).
84. M. Kasuga, T. Satoh, N. Tsukada, T. Yamazaki, F. Ogawa, M. Suzuki, I. Horikoshi, and T. Itoh: J. Soc.
Precision Eng., 57, 63 (1991).
85. J. Cummings and D. Stutts: Design for manufacturability of ceramic components, Am. Ceram. Soc.
Trans., 147 (1994).
86. A. Kumada: Ultrason. Technol. 1(2), 51 (1989).
87. Y. Tomikawa and T. Takano: Nikkei Mech., (Suppl.), 194 (1990).
88. Y. Okamoto, R. Yoshida, and H. Sueyoshi: Konica Minolta Tech. Rep., 1, p. 23 (2004).
89. K. Uchino and T. Ishii: J. Ceram. Soc. Japan, 96, 863 (1988).
90. K. Uchino: Proceedings of 5th Int’l Workshop on Piezoelectric Mater. Appl., State College, PA (Oct.
6–10, 2008).
91. H. W. Kim, S. Priya, K. Uchino, and R. E. Newnham: J. Electroceram., 15, 27–34 (2005).
92. H. W. Kim, S. Priya, and K. Uchino: Jap. J. Appl. Phys. 45(7), 5836–5840 (2006).
93. H. W. Kim, S. Priya, H. Stephanau, and K. Uchino: IEEE Trans. UFFC, 54(9), 1851–1859 (2007).
94. K. Uchino and A. Vazquez Carazo: Proc. 11th Int’l Conf. New Actuators, Bremen, Germany, June 9–11,
2008, A3.7, pp. 137–140 (2008).
95. http://www.lightningswitch.com/

8 Electro-Optic Devices
The electro-optic (EO) effect, which is the change in refractive index with an external electric field
applied, is manifested in certain materials which show tremendous promise of applications such as
light valves, beam deflectors, and optical displays for optical communications, in conjunction with
solid-state laser chips and optical fibers. Compared with liquid crystal devices, the ceramic EO com-
ponents, in general, possess advantages in terms of their response speed (μs), particularly in falling
time, contrast ratio (102) and gray scale (16 scales), and their ability to withstand high-intensity
illumination. On the other hand, the present ceramic components require relatively high drive volt-
ages (1 kV) and production cost ($100). Therefore, the development of a simple mass-production
process, electrode configurations with a narrow gap, and improvement of material properties, will
be key factors in the wide commercialization of ceramic optical components.
8.1 ELECTRO-OPTIC EFFECT—REVIEW

Let us initially review the operation of a light shutter based on the second-order EO effect (Kerr
effect). Birefringence Δn is induced in a crystal of the appropriate type when an electric field E is
applied, according to
Δn = −(1/2) Rn3 E 2. (8.1)
where
R is the quadratic EO coefficient
n is the original refractive index of the crystal
When this sample is placed between crossed polarizers arranged at the 45° direction with respect to
the E direction, and light is transmitted through the system, as shown in Figure 8.1, the output light
intensity is represented by
I = I 0 A sin 2 [(πRn3 L /2λ )E 2 ], (8.2)
where
I0 is the incident light intensity
A is an equipment constant (including the surface reflection and absorption effects)
L is the optical pathlength (i.e., the sample length)
λ is the wavelength of the light (typically 400 ~ 1600 μm)
The voltage required for the first intensity maximum is an essential device parameter called the
half-wave voltage (see Figure 8.1b). The light intensity output changes sinusoidally with the square
of E; the intensity modulation occurs more quickly with increasing electric field E.
8.2 TRANSPARENT ELECTRO-OPTIC CERAMICS

Since the 1960s the nonlinear polarizability of ferroelectrics has been investigated rather exten-
sively, and various EO and optical parametric devices have been developed. However, problems still
235

Analyzer
Polarizer
Half-wave voltage
–45° Imax
45°
L
PLZT
0 100 200 300

(a) Electric field (b) Applied voltage (V)
FIGURE 8.1 (a) Fundamental construction of an EO light shutter. (b) Light intensity change with applied
voltage. Imax = I0 A.
remain in preparing high grade optically homogeneous single crystals and, hence, manufacturing
costs are generally high.
The polycrystalline microstructure of a ferroelectric ceramic can also exhibit the EO effect if it
is sintered to a pore-free state to make it transparent. Relaxor ferroelectrics are of special interest
for nonlinear optic applications because an extraordinarily large apparent EO Kerr effect can be
observed even when the material is in its so-called paraelectric state. This section describes the
fundamental EO properties of perovskite-type polycrystalline and single crystal ferroelectrics.
8.2.1 (PB,LA)(ZR,TI)O3
The most useful ferroelectric EO materials have traditionally come from the (Pb,La)(Zr,Ti)O3
system; they generally have good transparency in a wavelength range extending from the visible to
infrared, and exhibit optical anisotropy with an applied electric voltage. Figure 8.2 shows the phase
diagram of the (Pb1−xLa x)(Zr1−y Tiy)1−x/4O3 system, on which are indicated the EO effects manifested
for various phase regions. Notice that the valence of lanthanum ion (3+) in the A-site (2+) generates
the vacancy of the B-site, as expressed in the composition formula. This atomic vacancy accelerates
the atomic diffusion during sintering, leading to high density (higher than 99%) and good transpar-
ency. However, it is important to understand that this transparency is for human visible wavelength
range, not for ultraviolet range. The hatched area, so-called Pen-ferroelectic region (because the
area shape looks like a “pen nib”), exhibits enhanced secondary EO effect, including the most popu-
lar composition PLZT (9/65/35), which is located in between the ferroelectric rhombohedral and
tetragonal, paraelectric cubic, and antiferroelectric phases (i.e., morphotropic phase boundary).
PbZr03 (mol%) PbZr03 PbTi03

100 90 80 70 60 50 40 30 20 10 0
FE, Rhomb FE, Tet
Memory
10 1st EO effect 10
AFE
(mol%) La
2nd
EO
20 effe 20
PE, Cub ct
30 30
FIGURE 8.2 Relation between PLZT composition, crystal structure, ferroelectricity, and the resulting
EO effect.

Electro-Optic Devices 237
Material PLZT 9/65/35 (2nd) PLZT 7/62/38 (1st)
P–E
14
Δn (×10–3) 10
Δn–E
5
–20 –10 0 10 20 –20 –10 0 10 20

Electric field Electric field
E (kV/cm) E (kV/cm)
FIGURE 8.3 Polarization P and birefringence Δn as a function of electric field E for PLZT (9/65/35)
(Kerr effect) and (7/62/38) (Pockels effect) ceramics.
The PLZT solid solution exhibits both the Pockels (primary) and Kerr (secondary) EO effects,
depending on the composition. Examples of typical birefringence Δn vs. electric field E curves, in
comparison with the polarization P vs. E, are shown in Figure 8.3 for PLZT (9/65/35) and (7/62/38)
compositions. The EO coefficients of the PLZT system are much larger than the values in conven-
tional crystals such as LiNbO3 and (Sr,Br)Nb2O6 (SBN) (see Table 8.1), which means that the voltage
required for the EO shutter is much less for the PLZT.
Among the various PLZT compositions, (9/65/35) is of particular interest, because it exhibits the
largest EO effect (R = 9.1 × 10 −16 m2/V2) and is thus applicable for light shutters and optical displays.
However, care must be taken to control grain size. Figure 8.4 shows the grain size dependence of
TABLE 8.1
Pockels (1st) and Kerr (2nd) EO Coefficients for
Various Materials
Material r (×10−10 m/V)
Primary EO coefficient LiNbO3 0.17
Ba2(K0.9Na0.1)Nb5O15 0.52
KH2PO4 0.52
(Sr0.5Ba0.5)Nb2O6 2.10
PLZT 8/65/35 (GS = 10 μm) 5.23
PLZT 8/65/35 (GS = 3 μm) 6.12
R (×10−16 m2/V2)
Secondary EO coefficient KTa0.65Nb0.35O3 5.30
PLZT 9/65/35 (GS = 2 μm) 9.12
PLZT 10/65/35 (GS = 2 μm) 1.07

1.7 μm
R-coefficient (×10–16m2/V2)
g-coefficient (×10–2m4/C2)
10 0.2
6
R-coefficient 0.1
4 g-coefficient
By G. H. Heartling
2
1.0 2.0 3.0 4.0 5.0

Grain size (μm)
FIGURE 8.4 Grain size dependence of the EO coefficients, R and g, for PLZT 9/65/35.
the EO coefficients, R and g (defined as Δn = −(1/2)Rn3P2 and −(1/2)gn3P2, respectively) for PLZT
9/65/35.1 The samples were prepared by hot-press sintering, starting from coprecipitated PLZT
powders. The EO response is drastically decreased for samples with grain sizes below 2 μm, which
corresponds approximately to the critical grain size below which the sample loses ferroelectric
properties (maybe antiferroelectric state).2 Therefore, relatively large grain size is necessary to pro-
duce a reasonable EO effect. On the other hand, a significant decrease in fracture toughness or
durability occurs for particularly large grain size samples, probably due to the B-site vacancies in
the crystal structure. A normally sintered transparent PLZT ceramic, with an average grain size of
more than 6 μm, has a fracture toughness (mode I for tensile stress) of KIC = 0.9 M N m−3/2, which
corresponds roughly to durability for 108 cycles of repeated operation.2 This translates to about only
2 months when the PLZT is used for an image display (TV) driven at 30 Hz.
EXAMPLE PROBLEM 8.1

PLZT 10/65/35, with a cubic symmetry shows an EO coefficient (R11 − R12) of 1.1 × 10 −16 (m2/V2)]
and n 0 = 2.49. Calculate the half-wave electric field for a sample with L = 1 mm, when λ = 633 nm
light is transmitted perpendicular to the electric field. Refer to the sample setting in Figure 8.1a.
Hint
The half-wave voltage is calculated from
Γ y = (π/ λ ) n03 E32 ( R11 − R12 )L = π, (P8.1.1)
where Γy is the phase retardation (refer to Equation 1.18).
Solution
E3 = [ λ /n03 ( R11 − R12 ) L ]1/ 2
= (633 × 10 −9 /2.493 × 1.1 × 10 −16 × 1 × 10 −3 )1/2
= 6.1 × 10 5 [V/m] (P8.1.2)

8.2.2 PMN–PT
The development of new ceramic EO materials with higher fracture toughness and larger EO coef-
ficients suitable for image display applications is still necessary due to some problems in PLZT as
discussed in Section 8.2.1. The following conditions should be considered:
1. Ceramic transparency requires almost zero birefringence in the zero-field state (i.e., a
pseudocubic structure) to suppress light scattering
2. Large fracture toughness may be obtained in a sufficiently dense structure (i.e., ion vacan-
cies are not suitable)
3. A large EO effect is manifested by relaxor ferroelectrics because of large electrostriction
The Pb(Mg1/3Nb2/3)O3-PbTiO3 (PMN–PT) system, which is known as a superior electrostrictive (sec-

ondary effect) material with very high fracture toughness (KIC = 1.7 M N m−3/2) is a good candidate
for EO applications.3 Samples of the (1 − x)PMN–xPT system were prepared by hot-press sintering
of oxide mixtures. The Curie temperature increases gradually with PbTiO3 content, passing room
temperature around x = 0.12, and the crystal structure is rhomboderal in the region below x = 0.4.
Figure 8.5 shows the composition x dependence of optical transmittance (λ = 633 nm) of a 0.5 mm thick
sample from the (1 − x)PMN–xPT system. The transmittance is reduced drastically above x = 0.14,
probably due to scattering caused by the spontaneous birefringence originated from the rhom-
boderal distortion. The best transmittance 49% is still smaller than the 62% observed for PLZT.
This suggests that a more sophisticated powder preparation technique will be required for fabricat-
ing adequately transparent PMN–PT ceramics.
The refractive index n (λ = 633 nm) is plotted as a function of composition x in Figure 8.6, and
shows a small maximum around x = 0.12. The values around n = 2.6 are slightly larger than the
n = 2.49 of PLZT 10/65/35. The most interesting data come from EO measurements. Figure 8.7a and b
shows the EO R coefficient and its corresponding hysteresis for λ = 633 nm, respectively, plotted
as a function of composition x. The maximum EO R-coefficient of 22 × 10 −16 m2/V2 for x = 0.12 is
more than twice that of PLZT 9/65/35 (9.1 × 10 −16 m2/V2). The hysteresis, defined as an equivalent
coercive electric field obtained from the experimental Δn curve, increases drastically above x = 0.16
and, hence, samples in this region cannot be used practically.
(1–x)PMN–xPT
50
40
Transmittance (%)
30
20
10
0
0.08 0.10 0.12 0.14 0.16
Composition x
FIGURE 8.5 Transmittance of a 0.5 mm thick sample of (1 − x)Pb(Mg1/3Nb2/3)O3−xPbTiO3 (for λ = 633 nm).

2.8
2.7
Refractive index n
2.6
2.5
(1–x) PMN–xPT
2.4
0.08 0.10 0.12 0.14 0.16

Ti fraction x
FIGURE 8.6 Refractive index as a function of composition x for (1 − x)Pb(Mg1/3Nb2/3)O3−xPbTiO3 (for λ =

633 nm).
×10–16
Electro-optic coefficient l (m2/V2)
20
10
0
0.08 0.10 0.12 0.14 0.16
(a) Composition x
×103
100
80
Hysteresis (V/m)
60
40
20
0
0.08 0.10 0.12 0.14 0.16
(b) Composition x
FIGURE 8.7 Changes in the EO coefficient R (a) and the corresponding hysteresis (b) in (1 − x)Pb(Mg1/3Nb2/3)
O3−xPbTiO3.

T (°C)
Pb(Zn1/3Nb2/3)O3
× 10–3
1 145
3 2 151
3 157
4 165
1 2
3 5 170
4
6 184
5 7 205
Birefringence Δn
2 8 212
0
0 5 10 15 × 105
Electric field (V/m)
FIGURE 8.8 Birefringence Δn vs. electric field E response in the paraelectric phase (elevated temperatures)
of a Pb(Zn1/3Nb2/3)O3 single crystal.
The data indicate that the 0.88 Pb(Mg1/3Nb2/3)O3−0.12 PbTiO3 has the potential to be a better EO
ceramic than PLZT with high mechanical toughness. However, higher optical transmittance must
be achieved by optimizing the fabrication process.
8.2.3 PB(ZN1/3NB2/3)O3
Pb(Zn1/3Nb2/3)O3 is a relaxor ferroelectric which can be used in single crystal form. In order to under-
stand the enhancement mechanism in the EO effect in relaxor ferroelectrics, this single crystal was
studied.4 Figure 8.8 shows the birefringence Δn vs. electric field E relation for a Pb(Zn1/3Nb2/3)O3
single crystal in the paraelectric phase (i.e., at elevated temperatures above TC).4 The single crystal
sample was made by a flux method using excess PbO. The parabolic curve in the low field region
becomes a straight line in the high field region.
A possible phenomenological analysis of this peculiar phenomenon is based on the model that the
crystal is composed of coexisting ferroelectric and paraelectric phases.4 Suppose that the volume frac-
tion of the paraelectric phase x(T) is given by an accumulated Gaussian distribution with respect to tem-
perature, the birefringence Δn is estimated by the summation of the linear and quadratic EO effects5:
Δn = [1 − x(T )]n3 (r33 − r13 )E/2 + x(T )n3 R44 E/2, (8.3)
where
n is the refractive index
r and R represent the EO Pockels (primary effect) and Kerr (secondary effect) coefficients,
respectively

E=0 E // <111>
E = E1
E = E2
FIGURE 8.9 Domain reversal mechanism in Pb(Zn1/3Nb2/3)O3 at room temperature. The spindle-like rhom-
bohedral up-down domains with the stripe period around 5 μm exhibit dynamic reversal under an electric field
cycle, by changing the color from “black” to “white,” and vice versa. Note that the domain wall motions are
not independent, but are synchronous, indicating a strong cooperative interaction between microdomains.
Even if the experimental data can be represented phenomenologically, the actual situation may
not be so simple as this model predicts as x(T) is also a function of the applied electric field E.
A more realistic description is found in terms of a microscopic domain reversal mechanism.
Pb(Zn1/3Nb2/3)O3 has very small spindle-like domains (5 μm) with ambiguous boundaries arranged
perpendicular to the external electric field. When a field greater than 0.5 kV/mm is applied, the
domain walls within a certain region of the sample move together or synchronously, such that the
microdomains respond to the applied field in a cooperative manner (see Figure 8.9).6 It is noteworthy
that the stripe period of the dark and bright domains (corresponding to up and down polarizations)
will not be changed by domain reversal, and that each domain area changes under an AC external
field with zero net polarization at zero field. The relaxor crystal can be electrically poled easily when
an electric field is applied around the transition temperature, and depoled completely without any
remanent polarization. This is the basis of the large “apparent” secondary nonlinear effects such
as electrostrictive and EO Kerr phenomena, which occur without any hysteresis. This phenomenon
is analogous to so-called super-paramagnetic effect observed in ultrafine ferromagnetic particle
samples, and thus is called “super-paraelectric” effect.
8.3 BULK ELECTRO-OPTIC DEVICES

8.3.1 FERPIC
One of the earliest applications is Ferpic (ferroelectric picture memory device). Figure 8.10 shows
the principle of the Ferpic.7 Initially, a ferroelectric PLZT 7/65/35 ceramic plate is uniformly
DC-poled laterally (see Figure 8.10a). Then, storage is achieved by switching domains normal to
the PLZT plate at points corresponding to the image’s high-intensity regions. To switch domains, a
high-contrast transparency is placed in front of the Ferpic and illuminated (Figure 8.10b), creating
low-impedance regions in the photoconductive film. The writing voltage supply will then cause
switching in these regions only.
Viewing/reading the memorized image is accomplished by passing polarized light through the
Ferpic and an analyzer as shown in Figure 8.10c. When the polarizer and analyzer are parallel, the
regions with remnant polarization normal to the plate produce a bright image, and the other regions
produce a dark image.

Poling
electrode
Poling electrodes Photoconductive
film
Incident Ceramic
light plate
Transparent
–
Poling electrode
voltage
Writing
supply
+ voltage
supply
+ –
Ceramic SW
plate High-contrast SW
transparency
(a) (b)
Observer
Collimated
light source
Polarizer Ferpic Analyzer
(c)
FIGURE 8.10 Principle of Ferpic and manufacturing process: (a) Initial DC poling on the PLZT plate (paral-
lel to the plate). This process also corresponds to erasing process. (b) Writing process using a photoconductive
film (realignment of the polarization normal to the plate). (c) Reading process using a pair of parallel polar-
izers. The normal polarization regions create bright image.
8.3.2 EYE PROTECTION APPLICATION

Sandia National Laboratories designed PLZT goggles for the U.S. Air Force to provide thermal and
flash-blindness protection for aircraft personnel.8 The goggle is basically a transverse-mode shutter
using an interdigital surface electrode configuration, similar to that shown in Figure 8.11. Different
from the stereo TV application, the light shutter uses a pair of parallel-arranged polarizers, so that
the light should be shut off when a half-wave voltage is applied on the PLZT.
8.3.3 STEREO TV APPLICATION

PLZT eye glasses for stereo TV (see Figure 8.11) have been fabricated using the light shutter prin-
ciple.9 The lenses consist of a pair of optically isotropic PLZT (9/65/35) discs sandwiched between
two crossed polarizers. When zero voltage is present between the electrodes, light will not be trans-
mitted. The transmitted light intensity increases with increasing applied voltage, and reaches a
maximum when a phase difference (retardation) of 180° is induced in the PLZT disk (i.e., at the
half-wave voltage).
The principle is the same as a stereo image toy which you enjoyed as a kid. You wore a pair of
glasses of red and blue lenses for viewing a red and blue superposed image. Stereo TV images of
an object are taken by two video cameras corresponding to the two eyes and the signal from each

TV Cameras
Left Right
Left image Right image
PLZT glasses
FIGURE 8.11 A stereo TV system using a pair of PLZT glasses. (From Kumada, A. et al., Proc. Ferroelectr.
Mater. Appl., 2, 205, 1977.)
camera is mixed alternately to make a frame for the right and left eyes. When viewing, the right and
left PLZT shutters are triggered synchronously to each image frame, resulting in a stereo image.
8.3.4 TWO-DIMENSIONAL DISPLAYS

The current requirements for high definition TV are stringent, and several systems have been
commercialized. One of the promising devices is a projection-type TV utilizing one-dimensional10
or two-dimensional (2-D) PLZT displays.11 The operating principle for a projection TV utilizing
2-D PLZT displays is introduced in this section. The development of a simple mass-production pro-
cess and the design of electrode configurations with a narrow gap are the key factors in the produc-
tion of the PLZT displays. A design shown in Figure 8.12 produces a very bright image with small
cross talk-related problems and is easy to manufacture.
8.3.4.1 Fabrication Process of the 2-D Display

The fabrication process for the 2-D PLZT light valve array is outlined in Figure 8.13.11 Coprecipitated
PLZT 9/65/35 powders were mixed with organic solvent and binder and then formed into a green
sheet. Platinum internal electrodes were printed on the green sheets. The electroded sheets were
then laminated, with the electrodes alternating by 90° between sheets, under a pressure of 3000 psi.
On
Ground
Float
0 [V]
V [V]
Light
(a) (b)
FIGURE 8.12 A 2-D PLZT optical display: (a) a matrix segment and (b) detail of an activated portion of the
display.

PLZT CP Organic
Binder solvent
Powder
Milling
Electrode
(vertical)
Slurry
Electrode
Doctor blade casting (horizontal)
PLZT green sheet

Green sheet
Electrode printing
Lamination
Al2O3 crucible
Firing
O2 gas
Multilayer PLZT
Cutting
Polishing
External electroding PLZT sample
2-D display Atmosphere powder
FIGURE 8.13 Fabrication process for the 2-D PLZT optical display.
Improved
performance
Transmittance (%)
60
40
20
0 400 500 600 700 800

Wavelength (nm)
FIGURE 8.14 Transmittance of the PLZT. Yellowish color was eliminated by an improved process.
The laminated body was sintered in an oxygen-controlled atmosphere, and the bulk was cut and pol-
ished. Finally the external electrodes were applied for vertical and horizontal addressing for indi-
vidual pixels on the display. The PLZT ceramic is transparent with slight yellowish color, as shown
in Figure 8.14. By optimizing the sintering condition (oxygen, Pb atmosphere, and temperature),
more square-like (dashed line in Figure 8.14) transmission curve was realized.

Vertical electrode External

Picture electrode
element
V-1 V-2 V-3 V-4 V-5 V-6 V-7 V-8 V-9 V-10
H–1
H–2
Horizontal electrode
H–3
H–4
H–5
10 mm
H–6
H–7
H–8
H–9
H–10
1 mm
20 mm
1.0 mm
(a) Front view Side view
10 mm
(b)
FIGURE 8.15 (a) Schematic electrode configuration of a (10 × 10) matrix PLZT light valve. (b) Top view
photograph of a PLZT light valve array with external electrodes.
Figure 8.15a shows the electrode configuration of a (10 × 10) matrix light valve. The shaded
portion of the device in the figure represents one image unit (pixel). The separated internal elec-
trodes are connected by external electrodes printed on the surface of the device. The continuous
(plate-through) electrodes are embedded 100 μm below the optical surface to avoid shorting with
the surface electrodes connecting the separated internal electrodes. Figure 8.15b shows a picture of
an actual display. Note that the layer thickness is about 0.35 mm.
8.3.4.2 Characteristics of the Light Valve Array

The optical transmittance of the PLZT device fabricated by the tape-casting technique was 62%
at λ = 633 nm, which is comparable with 63% transmittance for the ideal bulk sample prepared by
hot pressing. The brightness for red, green, and blue light was measured as a function of applied
voltage (Figure 8.16), where the electrode gap was 0.45 mm.11 Note that the half-wave voltage
differs for these three lights. The contrast ratio, defined as the ratio of the brightness on a screen
with the application of the half-wave voltage to the brightness with zero volt applied (220 cd/m2/2.8
cd/m2), of about 80, is superior to the values for conventional cathode ray tubes (CRT) or liquid
crystal displays (LCD). The response time (both rising and falling) of a single pixel on the display
is less than 10 μs, which is rapid enough to drive this shutter array at a raster frequency comparable
to the conventional CRT.

Luminous flux of light source : 10,000 lm

Distance from the projector : 1,500 mm
200 Red
Brightness (cd/m2)
Green
100
Blue
0 100 200 300 400

Applied voltage (V) (electrode gap: 0.45 mm)
FIGURE 8.16 Brightness on a screen vs. applied voltage for red, green, or blue light. Note that the half-wave
voltage differs for these three lights.
8.3.4.3 Construction of the Image Projector

The driving circuit for the display is shown schematically in Figure 8.17a. −80 V was continuously
applied on the horizontal electrodes from H-1 to H-10 with a time interval of 1 ms. +80 V, which was
a signal voltage, was applied on the vertical electrodes with various time intervals (multiple of 1 ms).
Note that only the pixels with a superposed voltage 160 V [=80 V − (−80 V)] are turned on, and the
pixels with 0 V or ±80 V are not on, which makes it possible to address those that are turned on. When
the terminals of the device are addressed, as shown in Figure 8.17b, the 10 pixels addressed (H-4,
V-3), (H-4, V-4), (H-4, V-5), (H-4, V-6), (H-4, V-7); (H-5, V-5), (H-5, V-7); (H-6, V-5), (H-6, V-7);
and (H-7, V-7) are on with continuous 1 ms time lag from the left to right pixels on the screen,
leading to the image appearing in Figure 8.17c (letter “F”).11
Cross talk was monitored on the 2-D display using the setup shown in Figure 8.18 with mono-
chromatic light.3 The test was made by keeping one vertical terminal (separated electrode) ON
(i.e., ground) and applying high voltage to multiple horizontal terminals (continuous plate-through
electrodes) simultaneously. There are three different cross talk patterns: vertical, horizontal, and
oblique types; that is, light leakage observed at vertically, horizontally, and obliquely adjacent pixels.
The results are shown in Figure 8.19a through c for three different input combinations, where the
top and bottom figures indicate the light intensity in microwatt (μW) on the screen for the ON and
OFF state, respectively. The leakage light intensity associated with the vertical and horizontal cross
talk was 20%–30% and 10%–20% of the main peak intensity, respectively, which does not affect the
image contrast significantly. On the other hand, oblique-type cross talk causes nonnegligible leak-
age, up to 50% depending on the applied voltage and the number of continuous electrodes addressed
(horizontal address) (called “combination”-type cross talk). Modification of the internal electrode
configurations is necessary to eliminate the cross talk problem completely.
Figure 8.20a illustrates a projection-type color TV system developed in collaboration with Fujitsu
General, Japan using three PLZT light valves for red, green, and blue. Dichroic mirrors were used
to generate red, green, and blue light beams started from Xenon lamp. Figure 8.20b shows the
response time of the light valve, where the light intensity was measured for a step voltage (160 V).
Both rising and falling response were about 10 μs, which is much quicker than the response time of
liquid crystal displays. The response time of the PLZT light valve can be limited by the mechanical
resonance period, because the PLZT is piezoelectric, as well as EO. The resonance frequency of
a pixel with a pair of 0.45 mm thick ceramic layers, which consist of 0.99 mm bulk surrounded by
neighbour pixels, can be estimated as a couple of 100 kHz. The measured response time of 10 μs
seems to be reasonable from this estimation.

Memory FET
Clock
V V V VV V V V V V
H–1 1 2 3 4 5 6 7 8 9 10
H–2
H–3
H–4
H–5 PLZT
Counter Decorder FET H–6 light valve
H–7
H–8
(a) H–9
H–10
1 ms
0V
H-1 10 ms –80 V
H-2
H-3
H-4
H-5
H-6
H-7
H-8
H-9
H-10
+80 V
V-3
0V
V-4
V-5
V-6
V-7
(b) V-1.2.8.9.10
(c)
FIGURE 8.17 (a) Driving circuit for the 2-D display. (b) The driving signal waveforms of the driving signal.
(c) An image of “F” on a screen illuminated through the PLZT projector.

Screen
Lens 2
Filter
Lens 1
Photodetector
Xenon lamp
Display
Slit
FIGURE 8.18 Cross talk test system. The light through a slit focused on the screen is measured.
Float
710
F 530
1720
Ground 690
660
Signal F 590
Cross talk F
0 0 V 0 0
Background
(a)
F
690 720
F 470 400
1610 1290
G 650 590
690 700
F 520 470
F
(b) 0 V V 0 0
F
690 780
F 600 570
1210 1900 980
G 590 560 610
890 720
F 430 480
F
(c) 0 V V V 0
FIGURE 8.19 Cross talk patterns for three different input combinations: (a) vertical type, (b) oblique type,
and (c) complex type.
EXAMPLE PROBLEM 8.2

In Figure 8.16, the first maximum in the light intensity is obtained at different voltages for red,
green, and blue light; 160 V for red, 150 V for green, and 130 V for blue.
1. Explain the reason physically.

2. Supposing that the refractive index n (= 2.49) and the EO coefficient (R11 − R12) [= 3.6 ×
10 −16 (m2/V2)] does not change significantly for each light, calculate the wavelength of
these three lights.

Mirror
Red Screen
PLZT light
valve
Dichroic mirror
(green ref.)
Green
Dichroic mirror
(blue ref.)
Blue Projection lens
Dichroic mirror
Xenon lamp (red ref.)
(a)
Current applied voltage
135 V
0.4 mA
Light intensity
10 μs
(b) Time
FIGURE 8.20 (a) Projection-type color TV system with using three PLZT light valves for red, green, and
blue. (b) Response time of the light valve.
Hint
The half-wave voltage is calculated from
Γ y = (π /λ )n03 E32 ( R11 − R12 )L
= π. (P8.2.1)
Solution
1. Since the half-wave voltage is provided by Equation P8.2.1, according to the illumination
light wavelength, the required voltage differs, as formulated in Equation P8.2.2:
λ = n03 E32 ( R11 − R12 )L. (P8.2.2)
Thus, for a shorter wavelength, a smaller electric field is required.

2. Taking into account the electrode gap of 0.45 mm, E3 = 3.55, 3.33, and 2.89 × 105 [V/m] for
R, G, and B, respectively, and a pathlength L given by (1.0–0.1) mm (note that the surface
depth 0.1 mm is an inactive layer):
λ = 2.493 × (3.55 × 10 5 )2 (3.6 × 10 −16 )0.9 × 10 −3
= 630 [nm] (for red)
λ = 555 [nm] (for green)
λ = 418 [nm] (for blue) (P8.2.3)
8.3.5 KTN SINGLE CRYSTAL OPTICAL SWITCH

K(Ta,Nb)O3 (KTN) is a well-known EO material. However, it has not been used widely so far due to
high preparation cost in a single crystal form and its mechanical fragility. Recent improved prepa-
ration technology made it possible to prepare single crystals of KTN as large as 40 × 40 × 20 mm.
In addition to this, their low cleavage characteristic even in a single crystal form makes it easy to
prepare balls/lenses by polishing.12 Figure 8.21 shows KTN balls prepared by polishing.
Fujiura et al. developed a K(Ta,Nb)O3 single crystal optical switch module, shown in Figure
8.22.12 Figure 8.23 shows its switching performance in C-band (1530–1570 nm).
1 mm
FIGURE 8.21 Low cleavage of a single crystal of K(Ta,Nb)O3, which makes it easy to prepare balls/lenses
by polishing. (From Fujiura, K., Bull. Ceram. Soc. Jpn., 41, 678, 2006. With permission.)
Bias Signal Temperature

voltage input control
PLC PLC
KTN
Input fiber Output fiber
FIGURE 8.22 KTN optical switch module. (From Fujiura, K., Bull. Ceram. Soc. Jpn., 41, 678, 2006. With
permission.)

0
Bar Cross
Attenuation (dB)
–10
–20
1530 nm 1540 nm 1550 nm

1560 nm 1570 nm
–30
0 1 2 3 4
Signal voltage (V)
FIGURE 8.23 Performance of the optical switch module in C-band (1530–1570 nm) made of K(Ta,Nb)O3.
(From Fujiura, K., Bull. Ceram. Soc. Jpn., 41, 678, 2006. With permission.)
y z
Image plane
z x
Polarizer Lens Electro-optic scanner Lens Piezoelectric scanner
FIGURE 8.24 Layout of hybrid scanner geometry. Deflection in the x-direction is due to the EO scanner
and deflection in the z-direction is due to the piezoelectric scanner. (From Scrymgeour, D.A. et al., J. Lightw.
Technol., 23, 2772, 2005. With permission.)
8.3.6 LASER BEAM SCANNER

A Penn State group developed a 2-D laser beam scanner by using a combination of EO and mechani-
cal mechanisms.13 The layout of this hybrid scanner geometry is illustrated in Figure 8.24. Deflection
in the x-direction is due to the EO scanner and deflection in the z-direction is due to the piezoelectric
scanner. The scanning frequency of the EO scanner can be chosen at any low or high frequency,
while it is fixed at the mechanical resonance frequency of the piezoelectric vibrator.
Figure 8.25 visualizes the operation of the hybrid scanner device. Figure 8.25a demonstrates
seven spots of the EO scanner for voltages of −300 to 300 V in steps of 100 V. Figure 8.25b shows
combination scanning with the EO scanner for the seven spots in (a) with the piezoelectric scanner
at 50 V scanning at its resonance of 25.6 kHz. Finally, Figure 8.25c shows 2-D scanning with the EO
scanner at 50 Hz and piezoelectric scanner at 25.6 kHz. The axis shown in (c) is the same coordinate
system as in Figure 8.24.
8.4 WAVEGUIDE MODULATORS

8.4.1 LINbO3 WAVEGUIDE
Light waveguides can be fabricated by depositing a high refractive index layer on a substrate.
The principle of the waveguide is shown schematically in Figure 8.26.14 Like an optical fiber,
light tends to bend toward high refractive index side, so that light is confined in the narrow
high refractive index layer fabricated on the crystal. Nb-diffused LiNbO3 single crystals are

(a)
5°
4°
3°
2°
1°
(b) 0°
5°
4°
3°
z
2°
x 1°
(c) 0°
0° 0.5° 1°
FIGURE 8.25 The hybrid device in operation. (a) Seven spots of the EO scanner for voltages of −300 to
300 V in steps of 100 V. (b) Combination scanning with the EO scanner for the seven spots in (a) with piezo-
electric scanner at 50 V scanning at 25.6 kHz. (c) 2-D scanning with the EO scanner at 50 Hz and piezoelectric
at 25.6 kHz. (From Scrymgeour, D.A. et al., J. Lightw. Technol., 23, 2772, 2005. With permission.)
TE1 TE0
n3 Superstrate
n1 Film
Distance x
(a) n2 Substrate
(b)
n3 n2 n1
FIGURE 8.26 Diagrams of (a) slab and (b) graded-index waveguides. The wave functions for the TE0 and
TE1 modes are shown in the refractive index profiles. (From Lines, M.E. and Glass, A.M., Principles and
Applications of Ferroelectrics and Related Materials, Clarendon Press, Oxford, 1977.)
commonly used. Note that the higher atomic number ion should have the higher electron
density and refractive index. Two diagrams of (a) slab and (b) graded-index waveguide are shown
in Figure 8.26. Wave functions for the TE0 and TE1 mode are also shown in the refractive index
profiles.14
Figure 8.27a and b are typical planar and ridge-type EO waveguides.15 Although fabrication of
a planar type is easy, the nonuniform distribution of the applied electric field is a problem. On the
other hand, as you can imagine, the ridge type requires a sophisticated manufacturing technology,
but the device function is close to the ideal. The transmitted light intensity is easily modulated by
applying a relatively low voltage. Phase modulation by 1 rad can be achieved by applying a voltage
of 0.3 V with power consumption of several microwatts per megahertz (μW/MHz).

Electrode
Waveguide
Guided light
(a) Substrate beam
c -axis
Ridge
Electrode
LiNbO3
crystal
(b)
FIGURE 8.27 EO waveguides: (a) planar type and (b) ridge type. (From Lines, M.E. and Glass, A.M.,
Principles and Applications of Ferroelectrics and Related Materials, Clarendon Press, Oxford, 1977.)
ITO top electrode
PZT waveguide layer
PLZT buffer layer
Nb:SrTiO3 semiconductor substrate 10.0 kV × 10.0 k 3.00 μm
(a) (b)
FIGURE 8.28 (a) Fundamental structure of an optical waveguide fabricated on a semiconductor substrate.
(b) SEM picture of the ridge-type PLZT thin film optical waveguide. (From Nashimoto, K., J. Ceram. Jpn.,
37, 538, 2002. With permission.)
8.4.2 PZT THIN FILM WAVEGUIDE

Various EO applications have been investigated by improving the film quality of PZT/PLZT. Figure
8.28a illustrates a fundamental structure of an optical waveguide fabricated on a semiconductor
substrate, and Figure 8.28b shows the SEM picture of the ridge-type PLZT thin film optical wave-
guide.16 Note that no columnar-like microstructure can be observed. A Y-type 1 × 2 optical switch
using a PLZT thin film waveguide was developed by Nashimoto et al. as shown in Figure 8.29a.16

3.0 mm
R = 8 mm
5 mm
127 mm
Electrode Channel waveguide
(a)
–3.0
y = –0.0139x –3.7327
R2 = 0.0167
–3.5
Intensity (dB)
–4.0
–4.5
0.0 0.5 1.0 1.5
(b) Propagation distance (cm)
FIGURE 8.29 (a) Y-type 1 × 2 optical switch using a PLZT thin film waveguide. (b) Optical transmission
loss characteristics of a slub-type PLZT thin film optical waveguide. (From Nashimoto, K., J. Ceram. Jpn.,
37, 538, 2002. With permission.)
By applying electric voltage, the refractive index below is decreased, leading to the laser beam
channel selection to a higher refractive index channel. Y separation length is one-half of the digital
switch using LiNbO3, because the EO coefficient is larger in PLZT. Figure 8.29b shows its optical
transmission loss of a slub-type PLZT thin film optical waveguide, demonstrating its high opti-
cal quality. Figure 8.30 demonstrates the voltage response of the optical switch. Under an applied
voltage of 10 V, −22 dB cross talk (contrast ratio) can be obtained, which is compared with several
tens volts in the conventional LN switch. The response speed (20 ns) shown in Figure 8.30b cor-
responds to the RC time constant.
Chapter Essentials
1. Relaxor ferroelectrics exemplified by PLZTs are widely applicable for EO light valve/
display applications. The superior characteristics of these materials are attributed primar-
ily to the easy poling of the ferroelectric microdomains (i.e., super-paraelectric effect).
2. An EO ceramic 0.88 Pb(Mg1/3Nb2/3)O3−0.12PbTiO3 with high mechanical toughness is one
of the promising new materials for long lifetime display applications.

On state path
0
–5
Output power (dB) –10
–15
Off state path

–20
–25
0 5 10 15 20
(a) Applied voltage (V)
Response = 20 ns
90%
10%
–250.00 ns 0.000 s 250.00 ns

(b) 50.0 ns/DIV Real time
FIGURE 8.30 (a) Voltage response of the Y-type 1 × 2 optical switch. Under an applied voltage of 10 V, −22
dB cross talk (contrast ratio) can be obtained. (b) Response speed demonstration. (From Nashimoto, K., J.
Ceram. Jpn., 37, 538, 2002. With permission.)
3. A PLZT 2-D light valve, fabricated by a tape-casting technique, is one excellent example
of a design well-suited to mass production at a low manufacturing cost.
4. A single crystal KTN optical switch is a recent development.
5. Light waveguides can be fabricated by depositing a high refractive index layer on a sub-
strate such as LiNbO3, or by using a PZT/PLZT thin film.

Check Point
1. (T/F) The refractive index is roughly determined by the electron density (per volume) of
the material. True or False?
2. We apply the electric field on cubic PLZT 9/65/35 along 3-axis direction. Does the refrac-
tive index for the polarized light along the 3-axis (i.e., extraordinary light) increase or
decrease with the electric field?
3. The randomly oriented light passed through a polarizer. When we neglect the light absorp-
tion by the polarizer, what percentage of the light intensity can we obtain from the polar-
izer, in comparison with the input intensity.
4. (T/F) The definition of Pockels EO coefficient r1jk is given by an expansion expression:
1/nij ( E ) − 1/nij (0) = Σrijk Ek . True or False?
5. Which PLZT composition is most popularly used for making a Kerr EO device, 7/62/38,
7/65/35, 8/65/35, or 9/65/35?
6. (T/F) The EO coefficient is increased by reducing the grain size below 1 μm. True or
False?
7. Let us consider PLZT transparent polycrystalline ceramic samples with almost 100% den-
sity (i.e., zero porosity) by a hot-press technique. When we prepare PLZT 9/65/35 (cubic)
and 9/40/60 (tetragonal), which will exhibit better transparency for the visible range?
8. Which color light has the lowest half-wave voltage, red, green, or blue, supposing that the
EO coefficient is almost the same for all three colors?
9. (T/F) A lower refractive index layer is fabricated on the LN single crystal substrate in order
to trap the light in this low index thin layer. True or False?
10. We fabricate two optical switches by using KTN 65/35 and PLZT 9/65/35, respec-
tively. The Kerr EO coefficients (R11 − R12) for KTN and PLZT are 5.3 × 10 −16 m 2/V2
and 9.1 × 10 −16 m 2/V2. Suppose that the same electric field is applied for both optical
switches, which device can reduce the optical pathlength (i.e., device thickness), KTN
or PLZT?
Chapter Problems
8.1 Let us consider a PLZT thin film (1 μm in thickness) deposited on a semiconductive
SrTiO3 substrate with the following two electrode configurations: (a) surface electrode
for lateral electric field and (b) surface electrode for normal electric field. Refer to
Figure 8.31.
Light direction
(a)
PLZT film
Light direction
(b)
FIGURE 8.31 Two electrode configurations for PLZT thin film optical waveguides: (a) surface electrode for
lateral electric field, and (b) surface electrode for normal electric field.

1. Discuss the merits and demerits of the above two-electrode configurations.

2. Suppose that the substrate is made of transparent single crystal and the electrode in
Figure 8.31b is made of a transparent material such as SnO2. Do you think the device
will work for light transmitted normal to the film?
Hint
1. Consider the birefringence and the shape of the optical indicatrix induced by the
electric field. Uniform electric field distribution is essential for precise control of
the light beam.
2. Consider the following retardation equation:
Γ y = (π /λ)n03 E23 ( R11 − R12 )L.
Is the pathlength L = 1 μm sufficient for obtaining Γy = π? How high electric field is

required in this case? Is this field realistic number achievable on the PLZT film sample?
8.2 Consider Kerr EO effect for a crystal with m3m symmetry.
1. Derive a secondary EO coefficient matrix for this symmetry.

2. Discuss the change in the optical indicatrix shape (refractive index ellipsoid), when
an electric field is applied along the z-axis.
3. Calculate the retardation, when the light is transmitted perpendicular to the electric
field, that is, along the y-axis. The pathlength is L (Figure 8.32).
Hint
1. The EO coefficient matrix is given as
⎛ R11 R12 R12 0 0 0 ⎞

⎜ ⎟
⎜ R12 R11 R12 0 0 0 ⎟
⎜ R12 R12 R11 0 0 0 ⎟
⎜ ⎟
⎜ 0 0 0 R44 0 0 ⎟
⎜ 0 0 0 0 R44 0 ⎟
⎜⎜ ⎟
⎝ 0 0 0 0 0 R44 ⎟⎠
2. Refractive indicatrix change of a cubit crystal with electric field. An original

sphere becomes a doughnut shape (Figure 8.33).
Polarizer Analyzer
45° –45°
L
z
PLZT
y
Electric field
FIGURE 8.32 Construction of an EO light shutter.

nz
y
ny
nx
x
FIGURE 8.33 Refractive indicatrix change of a cubic crystal under an applied electric field (Kerr effect).
REFERENCES
1. K. Tokiwa and K. Uchino: Ferroelectrics, 94, 87 (1989).
2. K. Uchino and T. Takasu: Inspection, 10, 29 (1986).
3. K. Uchino: Ceram. Int., 21, 309 (1995).
4. F. Kojima, J. Kuwata, and S. Nomura: Proc. 1st. Mtg. Ferroelectr. Mater. Appl. (Kyoto), p. 155 (1977).
6. R. Ujiie and K. Uchino: Proc. IEEE Ultrasonic Symp. (Hawaii), p. 725 (1990).
7. L. M. Levinson ed.: Electronic Ceramics, Marcel Dekker, New York, Chap. 7, p. 371 (1988).
8. J. T. Cutchen: Proc. 49th Annu. Sci. Mtg. Aerosp. Med. Assoc., New Orleans, LA, May (1978).
9. A. Kumada, K. Kitta, K. Kato, and T. Komata: Proc. Ferroelectr. Mater. Appl., 2, 205 (1977).
10. K. Murano: Ceram. Trans. Electro-Opt. Nonlinear Opt. Mater., 14, 283 (1990).
11. K. Uchino, K. Tokiwa, J. Giniewicz, Y. Murai, and K. Ohmura: Ceram. Trans. Electro-Opt. Nonlinear
Opt. Mater., 14, 297 (1990).
12. K. Fujiura: Bull. Ceram. Soc. Jpn., 41(9), 678–682 (2006).
13. D. A. Scrymgeour, B. Koc, L. Tian, M. W. Gentzel, K. Uchino, and V. Gopalan: J. Lightw. Technol.,
23(9), 2772–2777 (2005).
14. M. E. Lines and A. M. Glass: Principles and Applications of Ferroelectrics and Related Materials, p.
604, Clarendon Press, Oxford (1977).
15. I. P. Kaminov: Trans. IEEE, M. T. T., 23, 57 (1975).
16. K. Nashimoto: J. Ceram. Jpn., 37, 538 (2002).

9 PTC Materials
9.1 MECHANISM OF PTC PHENOMENON
9.1.1 PTC PHENOMENON
When barium titanate (BaTiO3, BT) is doped with lanthanum at levels less than 0.3 at.%, the ceramic
becomes semiconducting with a resistivity in the range of 10 ~ 103 Ω . cm. Moreover, the resistivity
is drastically increased, by 3 ~ 5 orders of magnitude, with increasing temperature around the Curie
point. This phenomenon was discovered in 1954, and is referred to as the PTC or PTCR (positive
temperature coefficient of resistivity) effect. Since then it has been investigated intensively by many
researchers.1,2 Figure 9.1 shows the impact of various dopants on the resistivity of the BaTiO3 PTC
ceramic as a function of temperature.
The PTC dopants typically have a higher ionic valence than either Ba (replaced by ions such as
La, Sm, Ce, or Gd) or Ti (replaced by ions such as Nb, Ta, and Bi) of the host structure. Since the
temperature at which the resistivity anomaly occurs is closely related to the Curie point, the tem-
perature coefficient can be easily modified by forming an appropriate solid solution with BaTiO3.
Figure 9.2 shows various PTCR curves for two such solid solution series. The solid line represents
BaTiO3 and the dashed lines show the effect of Sr (resistivity curve shifted to left) and Pb (resistivity
curve shifted to right) substitution in the proportions indicated.
9.1.2 MECHANISM OF THE PTC PHENOMENON

The theory for the PTC effect has not been established completely. Let us consider the effect in
terms of two mechanisms: the semiconducting properties of the doped BT and the grain boundary
barrier effect.
9.1.2.1 Semiconducting Properties

When sintered at high temperature, lanthanum-doped BaTiO3 is expected to become an n-type
semiconductor through the following reaction:3
Ba1− x La x TiO3 → Ba12+− x La 3x+ Ti14−+x Ti3x+ O23− .
The conduction takes place via transfer of electrons between titanium ions according to
Ti 4 + + e - ↔ Ti3+ .
Thus, the BaTiO3 grains in the ceramic are semiconducting, and remain semiconducting even on
cooling down to room temperature.
However, the grain boundary region changes during cooling. Oxygen is absorbed on the surface
of the ceramic, and diffuses to grain boundary sites, altering the defect structure along the grain
boundaries. The added oxygen ions attract electrons from nearby Ti3+ ions, thereby creating an insu-
lating barrier between grains. If excess oxygen ions are added per formula unit, the composition of
the grain boundary region can be described as follows:
(Ba 12−+x La 3x +)(Ti14−+x + 2 y Tix3−+2 y )O32+− y
261

100 M Mn 0.127
10 M
1M
Resistivity (Ω cm)
Cr 0.37
100 K
No addition
10 K
1K Si 1.0
100
10
100 200 300
Temperature (°C)
FIGURE 9.1 Resistivity as a function of temperature for several doped BaTiO3 PTCR ceramics. Dopant
concentrations are indicated near each curve.
108
r
+70%S
107
+ 60%S
106
×10 3
r
+ 50%S
105
ρ (Ω cm)
r
+ 40%S
104
%Pb
%Pb
Pb
+ 20 Sr
×10 3
+ 10%Pb
%Sr
Pb
Pb
+ 5%Pb
0%
0%
0%
8aTiO3
0%
+ 20
+ 15
+3
+4
+5
103
+3
102
101
–150 –100 0 + 100 + 200 + 300 + 400 + 450
T(°C)
FIGURE 9.2 Resistivity vs. temperature curves for BaTiO3 PTCR thermistors containing isovalent substitu-
tion of Sr or Pb for Ba.
This situation is illustrated schematically in Figure 9.3. Remember that the PTCR ceramics may
seem to be a composite consisting of semiconducting grains and highly resistive grain boundaries
in a 1:3 composite type.

PTC Materials 263
Grain
Barium titanate grain boundary
Ba O Ba O Ba O
O
O Ti O Ti O Ti O O
Oxygen
Ba O Ba O La O
O Ti O Ti O Ti O
Ba O La O Ba O
Electrons
FIGURE 9.3 Schematic illustration of the Ba1−xLa xTiO3 structure near the surface of a grain boundary.
Atmospheric oxygen dissociates and diffuses along a grain boundary, where the atoms attract electrons and
form insulating layers. (From Newnham, R.E., J. Mater. Educ., 6(5), 806, 1984.)
–eφ Conduction band

+ +
– – – + + – – –
+ +
L
Fermi level
NS
Grain boundary
FIGURE 9.4 Energy-level diagram near a grain boundary of the PTCR BaTiO3.
9.1.2.2 Schottky Barrier Effect

In order to explain the PTC or PTCR phenomenon, the most acceptable model is illustrated in
Figure 9.4, which was initially proposed by Heywang et al.1 When the two semiconductive (n-type)
ceramic particles are in contact at a grain boundary, an electron energy barrier (Schottky barrier) is
generated, and the barrier height is given by the following equation:
φ = eN S2/2ε0εN D , (9.1)
where
ND is the concentration of donor atoms
NS is the surface density of negatively charged acceptors (here assumed to be confined to the
surface due to Ba vacancies)

Note that the permittivity ε obeys the Curie–Weiss law:
ε = C/(T − T0 ) (9.2)
above TC. The low resistance at TC is thus accounted for by the lowering of the potential barrier due
to the increase in permittivity, as the temperature falls from an elevated temperature down to TC.
Below TC, the permittivity falls, but the spontaneous polarization appears and controls the electron
concentration to reduce the barrier height. This keeps the resistivity in a rather low range.
EXAMPLE PROBLEM 9.1

Electronic properties in ceramics are strongly affected by the surface layer or by the grain bound-
ary. Suppose that a grain boundary between n-type semiconductive grains possesses acceptor
impurities, and that electrons flow into the acceptor levels and an energy barrier is generated as
shown in Figure 9.4. Using the simple charge distribution model represented in Figure 9.5
(
ρ( x ) = eN D 0 < x < L ; ) ρ( x ) = 0 ( x > L) (P9.1.1)
answer the following questions:
(a) Describe the potential φ(x) by using the donor density ND, the barrier thickness L, the per-
mittivity ε0ε, and electronic charge e. Suppose that the change in φ occurs within a region
0 < |x| < L.
(b) Describe the barrier thickness L generated by the donor density ND and the surface accep-
tor density NS. You may use the charge neutralization condition.
(c) In semiconductive BaTiO3, the permittivity is decreased significantly above the Curie
temperature (=130°C). Explain the resistivity change by considering the barrier height
−eφ0.
Solution
(a) Poisson’s equation is given by
∂ 2 φ /∂x 2 = −eN D /ε 0 ε (0 < x < L ) (P9.1.2)
Taking into account the boundary conditions:
φ( L ) = φ( − L ) = φ(∞), − φ′( L ) = E = 0 (P9.1.3)
Charge density ρ(x) eND

+ + + + + +
+ + + + + +
–L L x
NS
FIGURE 9.5 Charge density distribution near the grain boundary between n-type semiconductive grains.
The grain boundary region has a highly resistive property.

PTC Materials 265
with a general solution φ(x) = −(eND/2ε0ε)x2 + Ax + B, we obtain the particular solution:
( )
2
φ( x ) = φ(∞) − (eN D /2ε 0 ε) x − L . (P9.1.4)
The potential depth at x = 0 is thus obtained as
φ0 = −(eN D /2ε 0 ε) L2 . (P9.1.5)
(b) Considering charge neutralization, we obtain
2eN D L = eN S . (P9.1.6)
Then,
L = N S /2 N D . (P9.1.7)
(c) The energy barrier height −eφ0 is represented as
−eφ0 = e2 N D L2 /2ε 0 ε
= e2 N S2/8ε 0 εN D . (P9.1.8)
Considering the Curie–Weiss law
ε = C/(T − T0 ). (P9.1.9)
we obtain
−eφ0 = (e2 N S2 /8ε 0CN D )(T − T0 ). (P9.1.10)
The barrier height increases in proportion to temperature. Since the resistivity is proportional to
exp(−eφ0/kT), it increases drastically with temperature (μexp(1 − T0/T)) for T > TC.
When we consider the situation below TC, the permittivity falls, but the spontaneous polariza-
tion appears and controls the electron concentration to reduce the barrier height. This keeps the
resistivity in a rather low range. For a temperature region much higher than TC, the resistivity
falls because of the very high thermal energy (kT) of electron, which tunnels through the energy
barrier.
9.1.2.3 Experimental Verification

There still remain two issues on the PTCR mechanism: (1) Is PTCR effect observed only in BT-based
ceramics? (2) Is just one boundary/contact sufficient to realize the PTCR effect?
Nomura and Doi reported the PTCR effect in a ferroelectric Pb(Fe1/2Nb1/2)O3 ceramic.4 Thus,
there is a possibility for any ferroelectric ceramic to exhibit the PTCR effect. However, superior per-
formance, that is, the resistance jump in 4 ~ 6 orders of magnitude can be found only in BT-based
materials.
For the second question, Niimi clearly demonstrated the PTCR effect between only two large
grains, as shown in Figure 9.6. Various PTCR performances (A, B, and C) in Figure 9.6b are related
with the crystal orientation difference between these two grains.

3
A
log R (Ω cm)
2
B
1
C
0
–1
20 μm 0 100 200
(a) (b) Temperature (°C)
FIGURE 9.6 (a) Demonstration of the PTCR effect between only two large grains of a BT ceramic sample.
(b) Various PTCR performances (A, B, and C) are related with the crystal orientation difference between these
two grains. (From Niimi, H., Mater. Integr., 19, 72, 2006. With permission.)
9.2 PTC THERMISTORS

PTC thermistors are applicable not only for temperature-change detection, but also for active cur-
rent controllers. The thermistor, when self-heated, exhibits a decrease in the current owing to a large
increase in resistivity. Practical applications for these devices are found in Ref. [6]
1. Ceramic heater
2. Demagnetization for a color TV (vacuum tube type)
3. Motor starter for a refrigerator compressor
4. Protector for excess current/voltage
“Ceramic heaters” have been widely commercialized in panel heaters, electronic thermos bot-
tles, and hair dryers. Figure 9.7 shows a PTC honeycomb air heater for hair dryers and automotive
chokes, manufactured by NGK, Japan.
EXAMPLE PROBLEM 9.2

The resistivity vs. temperature characteristic of a BT PTC ceramic is shown in Figure 9.8. Taking
into account heat generation through Joule heating, discuss the current vs. voltage relationship
under a room temperature ambient condition qualitatively.
FIGURE 9.7 PTC honeycomb air heater for a hair dryer. (Photo courtesy of NGK.)

PTC Materials 267
106 1 V/mm
105
B,C
PTC
ρ (Ω-cm)
104 magnitude
103
TC
102
ρ0 A
50 100 150 200

T(°C)
FIGURE 9.8 Resistivity vs. temperature characteristic of a BT PTC ceramic.
Solution
1. At the initial stage, the current–voltage relation obeys Ohm’s law (i.e., ρ is almost con-
stant), and power is dissipated via the Joule heating (V2/R) (Figure 9.9).
2. Around the point A (around Curie temperature 110°C), where a dramatic increase in ρ is
observed), the current becomes maximized, beyond which it decreases with increasing
applied voltage (see Figure 9.9).
3. Between the points B and C, the temperature of the device is almost stabilized, leading
to the relation V × I = constant. Notice that this is called “negative resistance or (∂I/∂V)”
meaning the current decreases with increasing the applied voltage.
4. Much above the point C (or above the resistivity peak, 200°C), a dramatic increase in cur-
rent is anticipated, because the saturation of the resistivity and the subsequent NTC (nega-
tive temperature coefficient) effect occur in this temperature range. Thus, to realize a stable
temperature in the PTC thermistor, the applied voltage must be adjusted between B and C.
Though the PTCR thermistor exhibits the 4 ~ 6 orders of magnitude change in the resistance, and
is applicable to many places, the minimum resistivity 50 Ω cm, or the actual resistance 50 Ω is still
a major problem. For improving the efficiency of the electronic equipment, this range of resistance
seems to be too high for the switch application viewpoint. Recently, a multilayer (ML)-type PTCR
thermistor was developed in order to reduce the minimum resistance down to 0.2 Ω.5 The technol-
ogy includes two key points: (1) grain size reduction and (2) thin layer thickness between a pair of
electrodes. Figure 9.10 shows grain size difference between the conventional PTCR ceramic and the
ML thermistor. Compared with a grain size larger than 5 μm in the conventional process, the ML
process dramatically suppressed the grain growth much less than 1 μm, which improved the PTCR
performance, as well as the performance reproducibility in the ML structure.
The actual difference between the conventional bulk PTCR ceramic and the ML thermistor is
shown in Figure 9.11.5 Notice that the resistance observed in ML type is 100 times lower than the
bulk type.

B
Current
Vs VMax
Voltage
FIGURE 9.9 Current vs. voltage relationship for a BT PTC ceramic. Notice the negative resistance region
between B and C points.
(a) 5 μm (b) 5 μm
FIGURE 9.10 Grain size difference between the conventional PTCR ceramic (a), and the ML thermistor (b).
(From Niimi, H., Mater. Integr., 19, 72, 2006. With permission.)
1,00,000
Bulk
10,000
1,000
Resistance (Ω)
100 ML type
10
1/100
0.1
0 100 200
Temperature (°C)
FIGURE 9.11 PTCR effect difference in the conventional bulk ceramic and the ML thermistor. (From
Niimi, H., Mater. Integr., 19, 72, 2006. With permission.)

PTC Materials 269
9.3 GRAIN BOUNDARY LAYER CAPACITORS

When a semiconductive BaTiO3 ceramic is oxidized to make a resistive surface layer, it can be used as
a high capacitance condenser. The capacitance is adjustable in the range of 0.4 ~ 0.5 μF/cm2. A new
type of grain boundary layer (GBL) capacitor has been developed using electrically resistive grain
boundaries on the semiconductive grains. The model for this structure is illustrated in Figure 9.12.
In practice, CeO2 or Bi2O3 is coated on a semiconductive ceramic, and diffused into the grain
boundaries by thermal treatment so as to make the boundary layer highly resistive. The resistive
GBLs of 1 μm thickness are fully connected in the ceramic with grain size of 10 μm. In this model,
the semiconductive grains are expressed by resistors, while the highly resistive grain boundaries
are expressed by capacitors. Both resistors and capacitors are connected in parallel and in series.
Example Problem 9.3 shows the reader how to calculate the apparent dielectric constant for this
model. This type of capacitor exhibits excellent frequency characteristics, and can be used as a wide
bandpass capacitor up to several GHz.
EXAMPLE PROBLEM 9.3

When a GBL capacitor is composed of many cubic core-shell units of a grain size D with a resis-
tive skin of dielectric constant εS, calculate the apparent dielectric constant εapp of this composite
material. Assume the sample has an electrode area S and an electrode gap d, and zero resistivity
inside the grains. The skin thickness δ is assumed to be equal to half of the grain boundary thick-
ness (from the calculation simplicity viewpoint).
Solution
As shown in Figure 9.13, let us divide the sample into two regions: C1 and C2.
(a) C1: thin thickness capacitors connected in series and in parallel

Since this is in a series connection [number of (d/D) in thickness] and in a parallel connection
[number of (S/D2) in area] of a unit capacitor with area (D − 2δ)2 and thickness 2δ, total capaci-
tance should be
C1 = [ ε 0 ε S ( D − 2δ )2/2δ ] × (S/D 2 )/(d/D)
= ε 0 ε S (S/d )( D/2δ) (δ /D << 1) (P9.3.1)
FIGURE 9.12 Model of the GBL capacitor.

Area S
d
D
C1 C2
(Conductor included) (εS)
FIGURE 9.13 Core-shell model of a GBL capacitor. Inside core (dotted part) has zero resistance and outside
shell is a capacitor.
(b) C2: narrow long capacitors
Since the area and the thickness are provided by [S − (D − 2δ)2(S/D 2)] and d, respectively,
C2 = ε 0 ε S (S/d )[1 − (1 − 2δ /D)2 ]
= ε 0 ε S (S/d )(4δ /D) (δ /D << 1) (P9.3.2)
Thus
Ctotal = C1 + C2
= ε 0 ε S (S/d )( D/2δ ) (P9.3.3)
The apparent dielectric constant is provided by εS(D/2δ), in which the enhancement factor is pro-
vided by (D/2δ). This value will reach 100, when D = 5 μm and 2δ = 0.05 μm (see Figure 9.10a).
Chapter Essentials
1. When BaTiO3 is doped with lanthanum at levels less than 0.3 atom %, the ceramic becomes
semiconducting with a resistivity in the range of 10 ~ 103 Ω . cm.
2. PTC effect: the resistivity is drastically increased, by 3 ~ 5 orders of magnitude, with
increasing temperature around the Curie point.
3. The Curie temperature (PTCR onset point) can be modified by substituting Ba with Sr
(lower temperature) or Pb (higher temperature).
4. Applications of PTC ceramics:
(1) Thermistors: over-current/voltage protectors, starting switches for motors, automatic
demagnetization circuits for color TVs
(2) Ceramic heaters: panel heaters, electronic thermos bottles, hair dryers, automotive
chokes
5. GBL capacitors: capacitors made of semiconductive BaTiO3 ceramics with highly resistive
grain boundaries.
Check Point
1. Provide a full expression of PTCR.

2. (T/F) The PTCR effect is a special phenomenon observed only in BaTiO3. True or False?

PTC Materials 271
3. (T/F) The PTCR effect creates negative resistance (i.e., voltage increase gives current
reduction). True or False?
4. (T/F) GBL capacitors are composed of two phases: resistive BT cores and semiconductive
grain boundary shells. True or False?
5. (T/F) The positive slope in the PTCR effect in BaTiO3 can be observed in the ferroelectric
phase, but not in the paraelectric phase. True or False?
Chapter Problems
9.1 Explain the current vs. voltage relationship for ZnO varistors, and describe their
applications in comparison with the PTCR material.
9.2 Let us consider the mechanism of the PTCR thermistor applications for over-current/
voltage protectors, starting switches for motors, and demagnetization circuits for color
TVs.
(a) When we apply a high AC voltage on the PTCR ceramic suddenly (step-like), the
current response is monitored as shown in Figure 9.14. Explain this current change
in terms of the PTCR phenomenon.
(b) When this PTCR thermistor is connected in parallel with the inductor/motor as in
Figure 9.15, explain how this circuit protects over-current problem in the inductor/
motor.
Hint
(a) Immediately after the AC voltage application, the resistance of the PTCR ceramic
becomes low. Thus, high AC current is observed. However, this current generates
self-heating in the ceramic, and the resistance will increase drastically after some
time, leading to the reduction of current.
(b) The impedance of the inductor/motor is given by jωL. Compared to this value,
the resistance of the PTCR ceramic is small at room temperature and large at an
elevated temperature. Thus, immediately after the AC voltage application, most of
the current will pass through the PTCR device. The current in the inductor/motor
will increase gradually with the temperature rise of PTCR.
Current
Time
FIGURE 9.14 Impulse current change in a PTCR ceramic (after an AC voltage).
R L
FIGURE 9.15 Over-current/voltage protector with a PTCR ceramic.

REFERENCES
1. W. Heywang: J. Am. Ceram. Soc., 47, 484 (1964).
2. E. Andrich: Electr. Appl., 26, 123 (1965–1966).
3. R. E. Newnham: J. Mater. Educ., 6(5), 806 (1984).
4. S. Nomura and K. Doi: Jpn. J. Appl. Phys., 9, 716 (1970).
5. H. Niimi: Mater. Integr., 19(3), 72–76 (2006).
6. Murata Mfg. Comp. Catalog: Mysterious Stones.

10 Composite Materials
Piezocomposites, composed of a piezoelectric ceramic and a polymer, are promising materials
because of their excellent tailorable properties. The geometry for two-phase composites can be
classified according to the connectivity of each phase (1, 2, or 3 dimensionally) into 10 structures:
0-0, 0-1, 0-2, 0-3, 1-1, 1-2, 1-3, 2-2, 2-3, and 3-3. In particular, a 1-3 piezocomposite, or PZT-rod/
polymer-matrix composite is considered most useful. The advantages of this composite are high
coupling factors, low acoustic impedance, good match to water or human tissue, mechanical flex-
ibility, broad bandwidth in combination with a low mechanical quality factor, and the possibility of
making undiced arrays by simply patterning the electrodes. The acoustic match to tissue or water
(1.5 Mrayl) of the typical piezoceramics (20–30 Mrayl) is significantly improved when it is incor-
porated into such a composite structure, that is, by replacing some of the dense and stiff ceramic
with a less dense, more pliant polymer. Piezoelectric composite materials are especially useful for
underwater sonar and medical diagnostic ultrasonic transducer applications.
In this chapter, other types of composites based on piezoelectric ceramics are also introduced.
Piezopassive-dampers are comprised of a piezoelectric ceramic particle, polymer, and a carbon
black, which suppress the noise vibration more effectively than traditional rubbers. Another type
of composite with a magnetostrictive ceramic and a piezoelectric ceramic produces an intriguing
product effect, the magnetoelectric effect in which an electric field is produced in the material in
response to an applied magnetic field.
10.1 CONNECTIVITY
Newnham et al. introduced the concept of “connectivity” for classifying the various PZT:polymer
composite structures.1 When considering a two-phase composite, the connectivity of each phase is
identified; e.g., if a phase is self-connected in all x-, y-, and z-directions, it is called “3”; if a phase
is self-connected only in z-direction, it is called “1.” A diphasic composite is identified with this
notation with two numbers “m–n,” where m stands for the connectivity of an active phase (such
as PZT) and n for an inactive phase (such as a polymer). In general, there are 10 types of diphasic
composites: 0-0, 1-0, 2-0, …, 3-2, 3-3, as illustrated in Figure 10.1 (Note that there are two different
configurations for 3-2 and 3-3).
A 0-0 composite, for example, is depicted as two alternating hatched and unhatched cubes, while
a 1-0 composite has Phase 1 connected along the z-direction. A 1-3 composite has a structure in
which PZT rods (1-dimensionally connected) are arranged in a 3-dimensionally connected polymer
matrix, and in a 3-1 composite, a honeycomb-shaped PZT contains the 1-dimensionally connected
polymer phase. A 2-2 composite indicates a structure in which ceramic and polymer sheets are
stacked alternately, and a 3-3 composite is composed of a jungle-gym-like PZT frame embedded in
a 3-dimensionally connecting polymer.
Try Example Problem 10.1 to familiarize the connectivity concept.

Identify the connectivity of the following two-phase composites:
(a) Multilayer piezoelectric actuator (piezoceramic and electrode metal)

(b) Positive temperature coefficient (PTC) honeycomb heater (barium titanate ceramic
and air)
273

0-0 1-0 2-0 3-0
1-1 2-1 3-1 2-2
3-2 3-2 3-3 3-3
FIGURE 10.1 Classification of two-phase composites with respect to connectivity. Two configurations are
shown for 3-2 and 3-3. (From Newnham R.E. et al., Mater. Res. Bull., 13, 525, 1978.)
(c) Barrier-layer (BL) capacitor (semiconductive grains with an insulative boundary)

(d) Steel-reinforced concrete (concrete with steel rods)
(a) (b) (c) (d)
Solution
(a) 2-2
(b) 3-1
(c) 0-3 (boundaries are connected 3-dimensionally)
(d) 3-1
10.2 COMPOSITE EFFECTS

There are three types of composite effects (Table 10.1): the sum effect, the combination effect, and
the product effect.
10.2.1 SUM EFFECTS

Let us discuss a composite function in a diphasic system to convert an input parameter X to an output
parameter Y. Assuming Y1 and Y2 are the outputs from Phases 1 and 2, respectively, responding to
the input X, the output Y * of a composite of Phases 1 and 2 could be an intermediate value between
Y1 and Y2. The figure in Table 10.1a shows the Y * variation with volume fraction of Phase 2 for a
case of Y1 > Y2. The variation may exhibit a concave or a convex shape, but the averaged value in a
composite does not exceed Y1, nor is it less than Y2. This effect is called a “sum effect.”

Composite Materials 275
TABLE 10.1
Composite Effects: Sum, Combination, and Product Effect
(a) Sum effect

Y1
Phase 1 : X---> Y1
Phase 2 : X ---> Y2 ] X ---> Y *
Y2
(b) Combination effect Phase 1 Phase 2
Phase 1 : X ---> Y1/Z1

Phase 2 : X ---> Y2/Z2 ] X ---> (Y/Z)*
Improvement
Y1
Y2
Phase 1 Phase 2 Y1/Z1 Y2/Z2

Z1
Phase 1 Phase 2
Z2
Phase 1 Phase 2
(c) Product effect
Phase 1 : X ---> Y
Phase 2 : Y ---> Z ] X ---> Z New function
An example is a fishing rod, i.e., a lightweight/tough material, where carbon fibers are mixed
in a polymer matrix (between 3-1 and 3-0). The density of a composite should be an average value
with respect to volume fraction, if no chemical reaction occurs at the interface between the carbon
fibers and the polymer, following the linear trend depicted in Table 10.1a. A dramatic enhancement
in the mechanical strength of the rod is achieved by adding carbon fibers in a special orientation,
i.e., along a rod (showing a concave relation as depicted in Table 10.1a).
Another interesting example is an negative temperature coefficient (NTC)–PTC material.2 V2O3
powders are mixed in epoxy with a relatively high packing rate (3-3), as illustrated in Figure 10.2.
Since V2O3 exhibits a semiconductor–metal phase transition at 160 K, a drastic resistivity change is
observed with increasing temperature. A further increase in temperature results in a larger thermal
expansion for epoxy than for the ceramic, leading to a separation of each particle, and the structure
becomes a 0-3 composite. The V2O3 particle separation increases the resistivity significantly at
around 100°C. Thus, the conductivity of this composite is rather high only over a limited tempera-
ture range (around −100°C to 100°C), which is sometimes called the “conductivity window.”
12.2.2 COMBINATION EFFECTS

In certain cases, the averaged value of the output, Y*, of a composite does exceed Y1 and Y2. This
enhanced output refers to an effect Y/Z which depends on two parameters Y and Z. Suppose that Y
and Z follow concave- and convex-type sum effects, respectively, as illustrated in Table 10.1b, the
combination value Y/Z will exhibit a maximum at an intermediate ratio of phases. This is called a
“combination effect.”

Resistivity (Ω·cm)
106
104
102
100
–100 0 100 200

Temperature (°C)
FIGURE 10.2 NTC–PTC effect observed in a V2O3:epoxy composite. (From Uchino, K., Solid State Phys.,
21, 27, 1986.)
Certain piezoelectric ceramic:polymer composites exhibit a combination property of g (the

piezoelectric voltage constant) which is provided by d/ε (d: piezoelectric strain constant, and ε:
permittivity). The details of these materials will be described in Section 10.3.
10.2.3 PRODUCT EFFECTS

When Phase 1 exhibits an output Y with an input X, and Phase 2 exhibits an output Z with an input Y,
we can expect for the composite an output Z with an input X. A completely new function is created
for the composite structure, called a “product effect.”
Philips developed a magnetoelectric material based on this concept.2 This material was composed
of magnetostrictive CoFe2O4 and piezoelectric BaTiO3 mixed and sintered together. Figure 10.3
shows a micrograph of a transverse section of a unidirectionally solidified rod of the materials with
an excess of TiO2 (1.5 wt.%). Four finned spinel dendrites CoFe2O4 are observed in cells (×100).
Figure 10.4 shows the magnetic field dependence of the magnetoelectric effect in an arbitrary unit
measured at room temperature. When a magnetic field is applied on this composite, cobalt ferrite
generates magnetostriction, which is transferred to barium titanate as stress, finally leading to the
generation of a charge/voltage via the piezoelectric effect in BaTiO3.
Since the magnetoelectric effect in a single phase material, such as Cr2O3, can be observed only
at a very low temperature (liquid He temperature), observation of this effect at room temperature is
really a breakthrough. Inexpensive sensors for monitoring magnetic field at room temperature or at
FIGURE 10.3 Micrograph of a transverse section of a unidirectionally solidified rod of mixture of magne-
tostrictive CoFe2O4 and piezoelectric BaTiO3, with an excess of TiO2 (1.5 wt.%).

ΔE
ΔH
Hmax Hdc
FIGURE 10.4 Magnetic field dependence of the magnetoelectric effect in a CoFe2O4:BaTiO3 composite
(arbitrary unit measured at room temperature).
elevated temperature can be produced from these composite materials. The recent development will
be discussed in Section 10.5.
10.3 PZT:POLYMER COMPOSITES

10.3.1 PIEZOELECTRIC COMPOSITE MATERIALS
As discussed in Section 7.1.2, polymer piezoelectric materials such as polyvinylidene difluoride
(PVDF) are very suitable for sensor applications. However, because of its small piezoelectric d
constants and very small elastic stiffness, PVDF cannot be used by itself in fabricating actuators or
high-power transducers. PZT:polymer composites, however, play a key role in the design of trans-
ducers, for applications such as sonar, which have both actuator and sensor functions.3
The piezoceramic:polymer composite was first reported by T. Kitayama et al. on the 0-3 type in 1972.3
After that, various studies have been done on the piezocomposites, as summarized in Table 10.2.
The representative data for several composite piezoelectric materials are listed in Table 10.3,4
with data for some single-phase piezoelectric polymers and PZT materials. The piezoelectric d
constant of PVDF, which indicates the strain per unit electric field (actuator applications), is 1/10
TABLE 10.2
Piezoceramic Composite Development History
Year Researchers Piezocomposite
1971 Kitayama (NTT, Japan) PZT powder:PVDF
1973 Powell (Gould) PZT powder:polyurethane
1976 Furukawa (Riken, Japan) Spherical Model
1978 Newnham (Penn State) Coral replica PZT:silicone rubber and
connectivity concept
1978 Seo (Mitsubishi Chemical, Japan) PZT powder:polyacetal
1980 Harrison (Honeywell) PZT powder:silicone rubber
1980 Newnham (Penn State) 1:3, 3:1, 3:2 types
1983 Banno (NTK, Japan) PT:polymer 0-3 type
1990 Uchino (Sophia Univ, Japan) PZT:carbon:polymer piezodamper
1990 Newnham (Penn State) Moonie/cymbal
2000 Uchino (Penn State) PZT:terfenol magnetoelectric composites
2004 Uchino (Penn State) Composite piezoenergy harvesting

TABLE 10.3
Comparison of the Piezoelectric Response of PZT:Polymer Composites, with the
Single-Phase Materials, PVDF and PZT
Piezoelectric Constants
Elastic Dielectric
Density Constant Constant d33 (10−12 g33 (10−3 gh (10−3m
Connectivity Material r (103kg m−3) c33 (GPa) e3 C N−1) mV N−1) V N−1)
— PZT(501A) 7.9 81 2000 400 20 3
Single phase
3-1 PZT:epoxy 3.0 19 400 300 75 40
3-3 PZT:silicone 3.3 3 40 110 280 80
rubber (replica
type)
PZT:silicone 4.5 19 400 250 60 —
rubber (ladder
type)
3-0 PZT:PVDF 5.5 2.6 120 90 85 —
PZT:rubber 6.2 0.08 73 52 140 30
PZT:chloroprene — — 40 — — 90
rubber
— Extended PVDF 1.8 3 13 20 160 80
Single phase
smaller than that of PZT, however, because of its small dielectric constant, the piezoelectric g
constant of PVDF, which indicates the voltage per unit stress (sensor applications), is 10 times
larger than that of PZT. PZT:polymer composites exhibit a wide range of piezoelectric response,
but in general d is slightly smaller than PZT and g is slightly smaller than PVDF. Thus, particu-
larly for underwater transducers, which perform both actuation and sensing and have a figure of
merit of d hgh, the composite materials are found to be far superior to single phase materials, like
PZT or PVDF.
10.3.2 PRINCIPLE OF PZT:POLYMER COMPOSITES

Here, in order to illustrate the principle, let us take a 1-3 composite which is composed of PZT fibers
embedded in a polymer matrix, as shown in Figure 10.5. The original fabrication process involves
the injection of epoxy resin into an array of PZT fibers assembled with a special rack.5 After the
epoxy is cured, the sample is cut, polished, electroded on the top and bottom, and finally electrically
poled. The die-casting technique has also been employed to make rod arrays from a PZT slurry.6
Piezoceramic fiber (phase 1)
3
2
Polymer matrix (phase 2)
FIGURE 10.5 A 1-3 composite of PZT rods and polymer. The top and bottom planes are rigid electrodes.

The effective piezoelectric coefficients d* and g* of the composite can be interpreted as follows:
When an electric field E3 is applied to this composite, the piezoceramic rods extend easily because
the polymer is elastically very soft (assuming the electrode plates which are bonded to its top and
* is almost the same as 1d of the PZT itself
bottom are rigid enough). Thus, d 33 33
*
d33 = 1d33 . (10.1)
Similarly
*
d33 = 1V 1d33 , (10.2)
where 1V is the volume fraction of phase 1 (piezoelectric). On the other hand, when an external stress
is applied to the composite, the elastically stiff piezoceramic rods will support most of the load, and
the effective stress is drastically enhanced and inversely proportional to the volume fraction. Thus,
larger induced electric fields and larger g* constants are expected:
*
g33 = d33
*
ε0ε*3 = 1d33 1V ε01 ε3
= 1 g33 1V . (10.3)
Figure 10.6 shows the piezoelectric coefficients for a PZT–Spurrs epoxy composite with 1-3 con-
nectivity, measured with a Berlincourt d33 meter. As predicted by the model for this composite,
the measured d 33* values are almost independent of volume fraction, but are only about 75% of
the d33 value of the PZT 501A ceramic. This discrepancy may be due to incomplete poling of the
rods. A linear relation between the permittivity and the volume fraction 1V is satisfied, resulting in
* with decreasing fraction of PZT. The piezoelectric coefficients for the
a dramatic increase in g 33
1-3 composite are listed in Table 10.3, together with those of a PZT–silicone composite with 3-3
connectivity. In conclusion, for the composites, the piezoelectric g coefficient can be enhanced by
two orders of magnitude with decreasing volume fraction of PZT, while the d coefficient remains
constant.
The advantages of this composite are high coupling factors, low acoustic impedance, good
matching to water or human tissue (more than 90% of a human body is water!), mechanical flex-
ibility, broad bandwidth in combination with a low mechanical quality factor, and the possibility of
making undiced arrays by simply patterning the electrodes. The thickness-mode electromechanical
400 2000
g33 80
300 1500 d33
d33 (×10–12 C/N)
g33 (×10–3 Vm/N)

Permittivity ε
60
200 1000
40
100 500
ε 20
0 0 0
0 10 20 30 40 50 60
PZT volume fraction (%)
FIGURE 10.6 Volume fraction dependence of the permittivity ε and the piezoelectric constants d33 and g33
in a 1-3 PZT:polymer composite.

coupling of the composite can exceed the kt (0.40 ∼ 0.50) of the constituent ceramic, approaching
almost the value of the rod-mode electromechanical coupling, k33 (0.70 ∼ 0.80) of that ceramic.7 The
acoustic match to tissue or water (1.5 Mrayl) of the typical piezoceramics (20 ∼ 30 Mrayl) is signifi-
cantly improved when they are incorporated in forming a composite structure, that is, by replacing
the dense, stiff ceramic with a low density, soft polymer. Piezoelectric composite materials are espe-
cially useful for underwater sonar and medical diagnostic ultrasonic transducer applications.
Although the PZT composites are very useful for acoustic transducer applications, care must be
taken when using them in actuator applications. Under an applied DC field, the field-induced strain
exhibits large hysteresis and creep due to the viscoelastic property of the polymer matrix. More
serious problems are found when they are driven under a high AC field, related to the generation of
heat. The heat generated by ferroelectric hysteresis in the piezoceramic cannot be dissipated easily
due to the very low thermal conductivity of the polymer matrix, which results in rapid degradation
of piezoelectricity.

A composite consists of two piezoelectric phases, 1 and 2, poled along the 3-axis and arranged
in a parallel configuration as shown in Figure 10.7a. Analogous to the terminology used in elec-
tronic circuit analysis, the structures pictured in Figure 10.7a and b are designed as “parallel” and
“series” connections, respectively. The volume fraction is 1V: 2V (1V + 2V = 1). Assuming that the
top and bottom electrodes are rigid enough to prevent surface bending, and that the transverse
piezoelectric coupling between Phases 1 and 2 is negligibly small in the parallel connection,
calculate the following physical properties of this composite:
(a) Effective dielectric constant ε*3

* coefficient
(b) Effective piezoelectric d 33
(c) Effective piezoelectric voltage coefficient g *33
Use the parameters D3, E3, X3, x3, and s E33 which are the dielectric displacement, electric field,
stress, strain, and the elastic compliance along the 3-axis (poling direction), respectively.
Solution
(a) Since the electrodes are common and E3 is common to Phases 1 and 2
D3 = 1V 1ε 3ε 3 + 2V 2 ε 3ε 0 E3
= ε*3ε 0 E3 . (P10.2.1)
Therefore
ε*3 = 1V 1ε 3 + 2V 2 ε 3 . (P10.2.2)
PZT (phase 1) Rigid electrode

Polymer (phase 2) Soft electrode
Poling
direction
(a) Parallel connectivity (b) Series connectivity
FIGURE 10.7 Diphasic composites arranged in (a) parallel and (b) series configurations.

(b) If Phases 1 and 2 are independent (no interaction)

1
x3 = 1d33 E3 , (P10.2.3)
2 2
x3 = d33 E3 . (P10.2.4)
However, when we assume that the top and bottom electrodes are rigid, the strain x3 must be
common to both Phases 1 and 2, and the average strain x *3 is given by the following equation:
1
V 1x3 − x3* ( ) 1
(
s33 = 2V x3* − 2 x3 ) 2
s33. (P10.2.5)
Therefore
⎣ (
x3* ⎡ 1V 2s 331d33 + 2V 1s 332 d33 ) (V s 1 2
33
⎦ )
+ 2V 1s 33 ⎤ E3 , (P10.2.6)
and consequently, the effective piezoelectric constant is given by
d3* (V s 1 2 1
33 33d + 2V 1s 332 d33 ) (V s 1 2
33 + 2V 1s 33 . ) (P10.2.7)
* = d * /ε ε*
(c) Since g 33 33 0 3
g33 = (V s
1 2
d + 2V 1s 332 d33
1
33 33 ) ⎣⎡( V s1 2
33 + 2V 1s 33 ε 0 ) (Vε 1 1
3 )
+ 2 V 2 ε3 ⎤ .
⎦
(P10.2.8)
10.3.3 THEORETICAL MODELS FOR 0-3 COMPOSITES

Various models have been proposed to predict the electromechanical properties of a composite
material. Pauer developed a 0-3 composite material comprised of PZT powder and polyurethane
rubber, and predicted its permittivity values by means of a cubes model.8 Figure 10.8 shows the
relative permittivity plotted as a function of volume fraction of PZT powder, in comparison with
theoretical values calculated on the basis of the cubes model (cubic PZT particles), the sphere model
Parallel model
1000
Relative permittivity ε33/ε0
Cube model
Sphere model
100 Measured
10
Series model
0 20 40 60 80 100
Volume fraction of PZT 1V (%)
FIGURE 10.8 Relative permittivity plotted as a function of volume fraction of PZT in PZT powder:
polyurethane rubber composites. Comparisons were made for the cube model, sphere model, and parallel and
series models.

Phase 2 Phase 1
(a)
a
(1– a)m
(1– a)n
a
(b) (1– a)1 a
FIGURE 10.9 Unit cell configuration for a 0-3 composite, according to Banno’s modified cubes model.
(spherical PZT particles), and the parallel and series models. Note that none of the models provided
a close fit to the experimental data.
Then, Banno proposed a “modified cubes model,” which took into account the anisotropic distri-
bution of cubes in x-, y-, and z-directions.9 The unit cell of this model is shown in Figure 10.9. The
following formulas can be derived for a uniaxially anisotropic case (i.e., l = m = 1, n ≠ 1):
( )
ε*33 = ⎡ a 2 a + (1 − a )n ⋅ 1ε 33 ⋅ 2ε 33 ⎤ ⎡⎣ a ⋅ 2ε 33 + (1 − a )n ⋅ 1ε 33 ⎤⎦
2
⎢⎣ ⎥⎦
( )
+ ⎡⎣1 − a 2 a + (1 − a )n ⋅ 2 ε 33 ⎤⎦ (10.4)
*
d33
⎣ ( )
= 1d33 ⎡ a 3 a + (1 − a )n ⎤ ⎡ a + (1 − a )n 1 ε33
⎦ ⎣ ( 2
) ( )
ε33 ⎤ ⎡(1 − a )n a + (1 − a )n + a 3 ⎤ ,
⎦ ⎣ ⎦
(10.5)
( )
= 1d31 ⎡ a 2 a + (1 − a )n ⎤ ⎡ a + (1 − a )n 1 ε33 ( ) (
ε33 ⎤ ⋅ a ⎡1 − a a + (1 − a )n ) + a3 ⎤ .
* 2 12
d31
⎣ ⎦ ⎣ ⎦ ⎢⎣ ⎥⎦
(10.6)
The volume fraction of Phase 1 is given by
1
(
V = a 3 a + (1 − a )n . ) (10.7)
The case n = 1 corresponds to the cubes model, and a general case 0 < n < 1 corresponds to a
configuration more dense along the z-direction. Figure 10.10 shows the experimentally determined
permittivity and piezoelectric d*h (=d*33 + 2d*31) coefficient for PbTiO3:chloroprene rubber compos-
ites, with the theoretical curves.10 When the volume fraction of PbTiO3 (1V) is small, n seems to be
less than 1 (i.e., the rubber thickness around a PbTiO3 ceramic cube is thinner along the z-direction
and thicker along the x- and y-directions) and with increasing the volume fraction, n approaches 1
(i.e., the rubber thickness becomes equal in all three dimensions). This configuration change may be
caused by the method of fabrication, which typically involves rolling and calendering.

500
Relative permittivity ε33/ε0

n=0
100
n = 0.1
50 n = 0.2
n = 0.3
n=1 m=0
10
5
0 0.2 0.4 0.6 0.8 1.0
(a) PZT volume fraction 1V
Piezoelectric dh (´10 –12 C/m)
100
n=0
n = 0.1
n = 0.2
n = 0.3
n=1 m=0
1
0.2 0.4 0.6 0.8 1.0
(b) PZT volume fraction 1V
FIGURE 10.10 Experimental values of the permittivity (a) and the piezoelectric d*h (=d*33 + 2d*31) coefficient
(b) for PbTiO3:chloroprene rubber 0-3 composites, shown with theoretical curves based on the Banno’s modi-
fied cubes model.
“Piezo rubber” composite
Outer insulating Center electrode wire

layer
Conductive rubber
+auxiliary wire
FIGURE 10.11 Structure of piezoelectric–rubber composite coaxial cables developed by NTK/NGK, Japan.
Figure 10.11 shows a structure of piezoelectric–rubber composite coaxial cables developed by

NTK/NGK, Japan, which were applied for the pick up of electric pianos and guitars, and automobile
speed monitoring on the highway.
10.3.4 ADVANCED PZT:POLYMER COMPOSITES

3-3 composites were first fabricated by the replamine method. A negative replica of a natural coral
structure with 3-3 connectivity was made of wax. Then, a positive replica of the negative structure was
prepared by introducing a PZT slurry into the porous network of the negative template, drying, burn-
ing out the wax, and finally sintering the PZT ceramic.11 In order to make highly porous PZT skel-
etons, the BURPS (BURned-out Plastic Spheres) method was proposed,12 where PZT powders and
plastic spheres are mixed in a binder solution, and the mixture is sintered. Miyashita et al. reported an
alternative method that involves piling up thin PZT rods in a 3-dimensionally connected array.13

Epoxy
PS PS
PZT
(a) (b)
FIGURE 10.12 3-s1 composites with (a) parallel and (b) series electrode configurations.
3-1 and 3-2 composites can be fabricated by drilling holes in a PZT block, and backfilled with
epoxy. In addition to this drilling method, an extrusion method has also been used to fabricate a PZT
honeycomb. The 3-1 and 3-2 composites show large dh and gh values.14 As shown in Figure 10.12,
there are two types of electrode configurations commonly applied to these composites: parallel [P]
and series [S]. In general, S types exhibit larger dh and gh values than P types do.
Materials Systems Inc. developed mass-production process for the 1-3 piezocomposites by using
injection molding, as schematically illustrated in Figure 10.13a and b.6 The photo of the piezocom-
posites is shown in Figure 10.14, compared with the U.S. quarter coin size.
10.4 COMPOSITE DAMPERS AND ENERGY HARVESTERS

The backgrounds of the piezoelectric dampers and energy harvesting devices were discussed in
Section 7.9. We review the same topics here from the composite application viewpoints. The com-
posite design is mandatory for improving the mechanical and electrical impedance matching, as
well as the electromechanical coupling enhancement.
10.4.1 PIEZOELECTRIC COMPOSITE DAMPERS

An intriguing application of PZT composites is a passive mechanical damper. Consider a piezo-
electric material attached to an object whose vibration is to be damped. When vibration is trans-
mitted to the piezoelectric material, the vibration energy is converted into electrical energy by the
piezoelectric effect, and an AC voltage is generated. If the piezoelectric material is in an open- or
short-circuit condition, the generated electrical energy changes back into vibration energy without
loss. The repetition of this process provides continuous vibration. If a proper resistor is connected,
however, the energy converted into electricity is consumed in Joule heating of the resistor, and the
amount of energy converted back into mechanical energy is reduced, so that the vibration can be
rapidly damped. Taking the series resistance as R, the capacitance of the piezoelectric material as
C, the vibration frequency as f, damping takes place most rapidly when the series resistor is selected
in such a manner that the impedance matching condition, R = 1/2πfC, is satisfied.15 Figure 10.15b
shows a mechanical damping difference for a unimorph-type beam (Figure 10.15a) under various
resistive shunt conditions. Notice that the resistance 6.6 kΩ (i.e., impedance matching condition)
realized the minimum period of the vibration.
Using this technique, in collaboration with ACX Company, K2 developed ski blades with PZT
patches to suppress unnecessary vibration during sliding.16 The integrated damping module in
Figure 10.16 measures 6.62 × 1.66 × 0.07 in., and includes piezoelectric wafers, energy dissipating
resistive shunt, and an LED function indicator. This module is allocated around 1/4 of the ski blade
length from the tip (i.e., the nodal point region).
The electric energy UE generated can be expressed by using the electromechanical coupling fac-
tor k and the mechanical energy UM:

1 Compound 2 Cold mold 3 Injection molding
PZT
powder
Thermo-
plastic
mix
Wax
4 Burn out binder

5 (low temp.)
Sinter
6
(high temp.)
Part solidified and ejected

(a)
Preform array
Voided PUR
matrix
Cast polyurethane
Demold
Blanchard grind to thickness

(b)
FIGURE 10.13 (a) Manufacturing process of the ceramic preform and (b) 1-3 composite fabrication steps,
developed by Materials Systems Inc. (From http://www.matsysinc.com/. With permission.)
UE = UM × k 2. (10.8)
The piezoelectric damper transforms electrical energy into heat energy when a resistor is con-
nected, and the transforming rate of the damper can be raised to a level of up to 50% (under the
electrical impedance matching condition). Accordingly, the vibration energy is decreased at a rate
of (1 − k2/2) times for a vibration cycle, since k2/2 multiplied by the amount of mechanical vibration
energy is dissipated as heat energy. As the square of the amplitude is equivalent to the amount of
vibration energy, the amplitude decreases at a rate of (1 − k2/2)1/2 times with every vibration cycle.
If the resonance period is taken to be T0, the number of vibrations for t s is 2t/T0. Consequently,
the amplitude in ts is (1 − k2/2)t/T . Thus, the damping in the amplitude of vibration in ts can be
0
expressed as follows:
(1 − k 2/2)t / T0 = exp (− t τ ), (10.9)

FIGURE 10.14 1-3 piezocomposites developed by Materials Systems Inc. (From http://www.matsysinc.
com/. With permission.)
Short 13 kΩ
Bimorph
1 kΩ 54 kΩ
Displacement (μm)
40
20
6.6 kΩ Open
0
0 40 80 120 160 0 40 80 120 160
Time (ms) Time (ms)
(a) (b)
FIGURE 10.15 Mechanical damping difference for a unimorph-type beam. (a) Unimorph beam structure,
and (b) damping performance under various resistive shunt conditions.
or
τ = −T0 ln(1 − k 2/2). (10.10)
In conclusion, the higher the k value is, the quicker the vibration suppression is.
Being brittle and hard, ceramics are difficult to assemble directly into a mechanical system.
Hence, flexible composites can be useful in practice. When a composite of polymer, piezoceramic

FIGURE 10.16 Piezoelectric damper patch (6.62 × 1.66 × 0.07 in.) with piezoelectric wafers, energy dissi-
pating resistive shunt, and an LED function indicator, developed by ACX and K2. (From ACX Company
catalogue: Passive Damping Ski. With permission.)
PZT ceramic Carbon Polymer
Piezoelectricity Conductivity Mechanical

flexibility
FIGURE 10.17 Piezoceramic:polymer:carbon black composite for vibration damping.
powder, and carbon black is fabricated (Figure 10.17), the electrical conductivity of the composite
is greatly changed by the addition of small amounts of carbon black.17 Figure 10.18 illustrates the
fabrication process of the damper composite. By properly selecting the electrical conductivity of the
composite (i.e., electrical impedance matching), the ceramic powder effectively forms a series cir-
cuit with the carbon black, so that the vibration energy is dissipated effectively. The conductivity of
the composite changes by more than 10 orders of magnitude around a certain carbon fraction called
the “percolation threshold,” where the carbon powder link starts to be generated. This eliminates the
use of external resistors. Note that the damper material exhibits a selective damping performance
for a certain vibration frequency, depending on the selected resistivity of the composite, which can
be derived from the electrical impedance matching between the permittivity and resistivity.
Figure 10.19 shows the relation between the damping time constant and the volume percentage
of carbon black in the PLZT:PVDF and PZT:PVDF composites. A volume percentage of about 7%
carbon black exhibited the minimum damping time constant, and therefore, the most rapid vibration
damping. Note that the PLZT with a higher electromechanical coupling k shows a larger dip (more
effective) in the damping time constant curve, which can be expected from Equation 10.10.

PVDF PLZT CB
Weighing
Rolling
Adding CB
and PLZT
to PVDF
Mixing
Press
FIGURE 10.18 Fabrication process of PLZT:PVDF composites containing carbon black.
200
τ/ms
150
PLZT
100 PZT
0 1 2 3 4 5 6 7 8 9
Volume percentage of CB
FIGURE 10.19 Damping time constant vs. volume percentage of carbon black in the PLZT:PVDF com-
posite. The minimum time constant (i.e., quickest damping) is obtained at the “percolation threshold.” (From
Suzuki, Y. et al., J. Ceram. Soc. Jpn., Int’l Edition 99, 1096, 1991.)
10.4.2 PIEZOELECTRIC COMPOSITE ENERGY HARVESTING

Energy recovery from wasted or unused power has been the topic of discussion for a long time.
Unused power exists in various forms such as vibration, flowing water, wind, human motion, and
shock wave. Flexible piezoelectric composites such as macro fiber composite (MFC) were employed

Interdigitated Structural epoxy

electrode pattern on
Inhibits crack
polyimide film (top and
propagation in ceramic.
bottom)
bonds actuator
components together.
Permits in-plane
poling and
actuation of
piezoceramic
(d33 vs. d31 Sheet of aligned
advantage) rectangular
piezoceramic fibers
Improved damage
tolerance and flexibility
relative to monolithic
ceramic.
FIGURE 10.20 MFC by Smart Material Corporation.
to the small force vibration sources where flexibility was necessary. The flexibility of PZT element
can generate the mechanical stress easily. In addition, it is easy to handle and shape the PZT ele-
ment. Previously, a PVDF film was used for implantable physiological power supply18 and inserted
for recovering some of the power in the process of walking.19 For the alternative, the advanced piezo
fiber composite (PFC) was experimentally demonstrated and confirmed for the energy conversion
component by the Penn State group recently.
Several flexible PZT composites such as active fiber composite (AFC) and MFC (Figure 10.20) are
commercially available, and used in high displacement actuator and transducer for energy conver-
sion. Major advantages of those fiber composites over conventional piezoelectric elements are their
flexibility and toughness, both far superior compared with monolithic PZT wafers. Furthermore,
due to their thin, planar geometry the fiber composites can be easily integrated into composite
laminates.20
The MFC is a revolutionary actuator which was developed at the NASA Langley Research
Center. The piezoelectric fibers manufactured by computer-controlled dicing saw and embedded in
the matrix were sandwiched between two layers of polyimide film that had a conductive interdigi-
tated electrode pattern printed on the inner surface. There are two types of MFC depending on the
polling direction. The poling direction of d33 type is parallel along fiber length and each segment has
opposite polling direction by interdigitated electrodes as shown in Figure 10.20. For the d31 type,
the poling direction is from the top to the bottom along the fiber thickness. The MFC is extremely
flexible, durable, and has the advantage of higher electromechanical coupling coefficients granted
through the interdigitated electrodes.
Previously, the d33 mode type of the MFC was tested to charge battery, but Sodano et al.21
claimed that the MFC did not produce high current because of the construction of the MFC.
The interdigitated electrodes of the d33 type made the small segments connect in series. Therefore
the voltage was added to generate high value but the current was not sufficient. Uchino et al. tested
the d31 mode type of MFC under small mechanical vibration source to generate enough current for
battery charging.22
The d31 mode type of MFC (M8528 P2) was fabricated by Smart Material Corp. The piezo-
ceramic fibers in the MFC are cut by 350 μm width and 170 μm thickness from piezoelectric
wafer by computer-controlled dicing saw. The total dimensions of MFC are 85 mm length,
28 mm width, and 0.3 mm thickness. Figure 10.21 shows the stress distribution of MFC when
the mechanical force is applied. As shown in Figure 10.21a, without any substrate, the exten-
sive stress and compressive stress occurred on the top and the bottom of MFC. In this case, the

MFC Substrate
Compression Compression
Extension Extension
(a) (b)
FIGURE 10.21 Stress distribution in the MFC: (a) without a substrate and (b) with an aluminum foil sub-
strate (0.17 mm in thickness).
neutral line of the stress distribution is in the middle of the MFC. For this reason the electrical
output will be very small. But in Figure 10.21b, when the additional substrate is bonded to the
bottom of the MFC, the neutral line will move down in the substrate. Therefore, the electrical
output will be increased because the MFC will have only a compressive stress or extensive stress
in whole volume. In addition, the thickness and material of the substrate should be carefully con-
sidered to keep the flexibility of the MFC. The minimum thickness of the substrate can be cal-
culated by the finite element method (FEM) simulation software, depending on the material to
escape the neutral line from the MFC. For this calculation, the dimensions of MFC are 85 mm ×
28 mm × 0.17 mm, and the thickness of a substrate is 0.17 mm. The aluminum foil was selected
to be used for the substrate in the experiment, even though the neutral line is located slightly
above the adjacent line.
The MFC was excited to generate a big bending motion by small force at the frequency of
1 ~ 5 Hz. Note that the small mechanical force used in the experiment means the minimum force
which can generate maximum strain in the flexible element without crack. The voltage signal of
the MFC is shown in Figure 10.22a. The voltage of MFC is considerably increased by bonding an
aluminum substrate. This signal was generated by a small mechanical force with a frequency of
5 Hz, and monitored by oscilloscope (TDS 420A). The aluminum substrate gave a lower flexibility
to the MFC, but the output voltage from MFC with small bending showed the much higher voltage
signal. The output voltage signal from the MFC was passed through the rectifier and charged to
the capacitor, and successively discharged through a resistive load. The rectified voltage and output
power was shown in Figure 10.22b. The generated electric power from a small mechanical force at
5 Hz was around 1.5 mW at 200 kΩ.
The MFC shows the possibility of the intelligent clothing (IC) with flexible piezoelectric energy
harvesting system, aiming at a general power source for charging up portable equipment such as
cellular phones, health monitoring units, or medical drug delivery devices.
10.5 MAGNETOELECTRIC SENSORS

Similar to nuclear radiation, magnetic irradiation cannot be easy felt by human. Some reports men-
tioned that brain cancer may be triggered by a frequent usage of a mobile phone, though there
is no strong scientific evidence. The problem is that we cannot even purchase a magnetic field
detector for a low frequency (50 or 60 Hz). The Penn State, in collaboration with Seoul National
University, developed a simple and handy magnetic noise sensor for environmental monitoring pur-
pose.23 Figure 10.23 shows a schematic structure of the device, in which a PZT disk is sandwiched
by two Terfenol-D (magnetostrictor) disks. When a magnetic field is applied on this composite,
Terfenol metal will expand, which is mechanically transferred to PZT, leading to an electric charge
generation from the PZT. By monitoring the voltage generated in the PZT, we can detect the mag-
netic field. Figure 10.24 shows the experimental result comparison of the magnetic field monitoring
performance of the devices with different PZT materials, which indicates that a high piezoelectric

Without substrate
With aluminum substrate
40
20
Output voltage (V)
–20
–40
–60
–3 –2 –1 0 1 2
(a) Time (s)
2.0
26
24 1.8
22
20 1.6
18
1.4
16
14
Rectified voltage (Vrec) 1.2
12
Output power (mW)
10 1.0
50 100 150 200 250 300 350 400 450 500 550
(b) Load (kΩ)
FIGURE 10.22 Output of the MFC by around 5 Hz. (a) Output voltage signal of the MFC and (b) recti-
fied voltage and output power with a resistive load. (From Uchino, K., Proc. 9th Jpn. Int’l SAMPE Symp.,
pp. 11–14, Nov. 18–21, Tokyo, 2005.)
g constant material (APC 840) exhibits the largest performance. The key of this device is highly
effective for a low frequency.
Chapter Essentials
1. Composite effects: (1) sum effect, (2) combination effect, (3) product effect.
2. PZT:polymer composites:
(a) High dhgh constant
(b) Good acoustic impedance matching with water and human tissue
(c) Mechanical flexibility

Permanent
magnet
Terfenol-D PZT
FIGURE 10.23 Schematic structure of the magnetoelectric device, in which a PZT disk is sandwiched by
two Terfenol-D (magnetostrictor) disks.
5000
APC 840
4000 PZT–5A
APC 841
dE/dH (mV/cm Oe)
3000
2000
1000
0 1000 2000 3000 4000 5000

Magnetic field (Oe)
FIGURE 10.24 Comparison of the magnetic field monitoring performance of the electromagnetic devices
with different PZT materials.
3. 1-3 composites:
The effective piezoelectric coefficients d* and g* are provided by
*
d33 = 1d33
*
d31 = 1V 1d31
*
g33 = d33
*
/ε 0 ε *3 = 1d33 /1V ε 01 ε *3 = 1 g33 /1V ,
where 1V is the volume fraction of Phase 1 (piezoelectric).
4. The principle of mechanical damping:

(1) Vibration is transmitted to the piezoelectric material.
(2) Vibration energy is converted into electrical energy (AC voltage) through the piezo-
electric effect.
(3) If a proper resistor is connected, the energy converted into electricity is consumed as
Joule heat by the resistor.
(4) The energy converted back into mechanical energy is reduced, so that the vibration can
be rapidly damped.
(5) Damping takes place most rapidly when the series resistor is selected in such a manner
that the impedance matching condition, R = 1/2πfC, is satisfied.

5. Elastically soft (flexible) piezoelectric composites such as MFCs are suitable to harvest low
electrical energy from human motion.
6. The principle of magnetoelectric composites with a combination of magnetostrictor
(Terfenol) and piezoelectric (PZT):
(1) Magnetic field (even low frequency) generates strain in Terfenol.
(2) Strain induced in Terfenol is transferred to PZT laminated as stress.
(3) PZT generates electric field via direct piezoelectricity.
(4) A high-g constant PZT is preferred for high sensitivity.
Check Point
1. Identify the connectivity (x–y) of a steel-reinforced concrete composite (concrete with steel
rods).
2. Which type of composite effect do you expect from a steel-reinforced concrete composite,
sum effect, combination effect, or product effect?
3. Which connectivity of PZT–polymer composites provide larger piezoelectric g constant,
parallel connectivity or series connectivity?
4. (T/F) In the 1-3 PZT:polymer composite, the effective permittivity of the composite ε*3
decreases on increasing the volume fraction of the PZT phase 1V. True or False?
5. (T/F) In the 1-3 PZT:polymer composite, the effective piezoelectric d constant of the
composite d*33 decreases on increasing the volume fraction of the PZT phase 1V. True or
False?
6. (T/F) In the 1-3 PZT:polymer composite, the effective piezoelectric g constant of the
composite g*33 decreases on increasing the volume fraction of the PZT phase 1V. True or
False?
7. Nominate three single materials to make an elastically flexible damping composite
(based on the piezoelectric passive damping principle) to be used for engine vibration
suppression.
8. (T/F) Because the human tissue is elastically soft, its motion energy is rather easily trans-
ferred to a soft piezoelectric material, in comparison with the case of a hard PZT. True or
False?
9. (T/F) Because the piezoelectric energy harvesting device has relatively small output elec-
trical impedance, it is easy to charge up a rechargeable battery just by connecting them.
True or False?
10. Nominate two single materials to make a two-phase composite, in order to create a mag-
netoelectric sensor at room temperature.
Chapter Problems
10.1 Two kinds of piezoelectric materials, 1 and 2, poled along the 3-axis compose a com-
posite in a series configuration as shown in Figure 10.7b. The volume fraction is 1V:2V
(1V + 2V = 1). Assuming that the top and bottom electrodes are soft enough not to prevent
surface extension, and that the transverse piezoelectric coupling between Phases 1 and 2
is considered to be just connected elastically, calculate the following physical properties
of this composite:
(a) Effective dielectric constant ε*3
(b) Effective piezoelectric d*33 coefficient
(c) Effective piezoelectric voltage coefficient g*33
Use the parameters D3, E3, X3, x3, s33 which are the dielectric displacement, the elec-
tric field, the stress, the strain, and the elastic compliance, respectively. Note that the

voltage/electric field applied on the piezoelectric phase will be drastically reduced with
sandwiching low permittivity polymer layers in series.
10.2 Derive the following equations from the “modified cubes model,” which takes into
account the anisotropic distribution of cubes with respect to the x-, y-, and z-directions,
as illustrated in Figure 10.9 (l = m = 1, n ≠ 1), where the volume fraction of Phase 1 is
given by
1
V = a 3/(a + (1 − a)n)
ε*33 = [ a 2 (a + (1 − a)n)2 ⋅ 1ε33 ⋅ 2ε33 ]/[ a ⋅ 2ε33 + (1 − a)n ⋅ 1ε33 ]
( )
+ ⎡⎣1 − a 2 a + (1 − a )n . 2 ε 33⎤⎦
*
d33 = 1d33[ a 3 (a + (1 − a)n)]/[ a + (1 − a)n(1 ε33 / 2ε33 )]
/ [(1 − a)n /(a + (1 − a)n) + a 3 ]

*
d31 = 1d31[ a 2 (a + (1 − a)n)]/[ a + (1 − a)n(1 ε33 / 2 ε33 )]
⋅ a / [1 − a(a + (1 − a)n)1/2 + a 3 ].
10.3 We will consider PZT:polymer composites as a pyroelectric sensor, rather than a piezo-
electric device. In addition to the primary pyroelectric coefficient α (=−(∂P/∂T)),
a secondary pyroelectric effect is anticipated in a composite structure due to a large differ-
ence in thermal expansion coefficients between PZT and a polymer. Discuss this secondary
pyroelectric effect for both parallel and series connections, as illustrated in Figure 10.25.
Assume that
1. In the parallel connection, the top and bottom electrodes are rigid enough to prevent
surface bending (constrained just along the up-down z-direction, but free along the
lateral direction), and that the transverse stress between Phases 1 and 2 is negligibly
small.
2. In the series connection, the top and bottom electrodes are negligibly soft, in com-
parison with Phases 1 and 2.
The volume fraction is 1V : 2V (1V + 2V = 1). Use the parameters T, αT, X3, x3, and s33
which are temperature, the thermal expansion coefficient, the stress, the strain, and the
elastic compliance, respectively.
PZT (phase 1) Rigid electrode

Polymer (phase 2) Soft electrode
Poling
direction
(a) (b)
FIGURE 10.25 Composite structures for a pyroelectric material. (a) Parallel connectivity. (b) Series
connectivity.

0.3
Piezoelectric
ceramics CFRP
50
60
90
90
t 0.3
CFRP Piezoelectric
20
ceramics
10
Clamp Clamp
0.65 15 0.65 15
FIGURE 10.26 Two configurations of CFRP/PZT cantilever beams; (a) a pair of PZT plates are bonded on
the tip part of the CFRP plate and (b) on the bottom part of the CFRP with a bimorph configuration. (From
Tanimoto, T. et al., Jpn. J. Appl. Phys., 36, 6110, 1997. With permission.)
10.4 Passive damping systems of carbon-fiber-reinforced plastic (CFRP)/PZT beams are

developed. Figure 10.26a and b shows two configurations to study; a pair of PZT plates
are bonded on the tip part of the CFRP plate, and on the bottom part of the CFRP
with a bimorph configuration, respectively. When we consider the initial three vibration
modes (1st, 2nd, and 3rd bending modes), discuss on which bimorph configuration (a)
or (b) is effective on which modes.24
Hint
Consider that the mechanical energy concentrated at the nodal range should overlap
with the bimorph part to effectively damp the vibration. Refer to Figure 10.27 with three
initial bending modes of the beam.
1st resonance mode Bimorph-A, B
2nd resonance mode Bimorph-A, B
3rd resonance mode Bimorph-A, B
FIGURE 10.27 Three initial bending modes of the beam. (From Tanimoto, T. et al., Jpn. J. Appl. Phys., 36,
6110, 1997. With permission.)

REFERENCES
1. R. E. Newnham et al.: Mater. Res. Bull., 13, 525 (1978).
3. T. Kitayama and S. Sugawara: Study Comm. Electron. Circuit Compon. Mater., Ref. no. CPM72-17
(1972).
4. K. Uchino, S. Nomura, and R. E. Newnham: Sensor Technol., 2, 81 (1982).
5. K. A. Klicker, J. V. Biggers, and R. E. Newnham: J. Am. Ceram. Soc., 64, 5 (1981).
6. http://www.matsysinc.com/, U.S. Patent 5,340,510.
7. W. A. Smith: Proc. IEEE Ultrasonic Symp. 89, p. 755 (1989).
8. L. A. Pauer: IEEE Int’l Convention Record, 1–5 (1973).
9. H. Banno: Proc. 6th Int’l Meeting on Ferroelectricity (IMF-6, Kobe, 1985); Jpn. J. Appl. Phys., 24(Suppl.
24–2), 445 (1985).
10. H. Banno and T. Tsunooka: Ceramic Data Book ‘87, Industrial Product Technology Soc., Tokyo, p. 328
(1987).
11. D. P. Skinner, R. E. Newnham, and L. E. Cross: Mater. Res. Bull., 13, 599 (1978).
12. T. R. Shrout, W. A. Schulze, and J. V. Biggers: Mater. Res. Bull., 14, 1553 (1979).
13. M. Miyashita et al.: Ferroelectrics, 27, 397 (1980).
14. A. Safari, R. E. Newnham, L. E. Cross, and W. A. Schulze: Ferroelectrics, 41, 197 (1982).
15. K. Uchino and T. Ishii: J. Ceram. Soc. Jpn., 96, 863 (1988).
16. ACX Company catalogue: Passive Damping Ski.
17. Y. Suzuki, K. Uchino, H. Gouda, M. Sumita, R. E. Newnham, and A. R. Ramachandran: J. Ceram. Soc.
Jpn., Int’l Edition 99, 1096 (1991).
18. E. Hausler, L. Stein, and G Harbauer: Ferroelectrics, 60, 277 (1984).
19. J. Kymissis, C. Kendall, J. Paradiso, and N. Gershenfield: Digest of Papers, Second IEEE International
Symposium on Wearable Computers, Pittsburgh, PA, p. 132 (1998).
20. L. J. Nelson: Mater. Sci. Technol., 18, 1245 (2002).
21. H. A. Sodano, G. Park, D. J. Leo, and D. J. Inman: Smart structure and materials: Smart sensor technol-
ogy and measurement systems, Proc. of SPIE, 5050, 101 (2003).
22. K. Uchino: Proc. 9th Jpn. Int’l SAMPE Symp., pp. 11–14, Nov. 18–21, Tokyo (2005).
23. J. Ryu, A. Vazquez Carazo, K. Uchino, and H. E. Kim: Jpn. J. Appl. Phys., 40, 4948–4951 (2001).
24. T. Tanimoto, K. Horiuchi, and K. Uchino: Jpn. J. Appl. Phys., 36 Part I (No. 9B), 6110 (1997).

11 Future of Ferroelectric Devices
We have studied the fundamentals and applications of ferroelectrics, including
1. High-permittivity dielectrics
2. Ferroelectric memories
3. Pyroelectric devices
4. Piezoelectric devices
5. Electro-optic devices
6. PTC materials
7. Composite materials
From a viewpoint of commercialization, capacitor dielectrics dominate at present, followed by

piezoelectric vibrators such as buzzers and speakers. Among the other classes of devices, sales are
relatively low.
What will be the next promising market for ferroelectric devices? As we have seen, ferroelectrics
can be utilized for various applications, but have failed to be commercialized in many cases in the
past. In the case of the light sensor, for example, semiconductive materials are superior to ferro-
electrics in response speed and sensitivity. Magnetic devices are popular for memory applications.
High-permittivity dielectric thin film can survive in DRAMs, but commercialization of ferroelec-
tric memory (FeRAM) is uncertain because of the variability of the coercive field of the material,
and flash memory devices dominate the market already. Liquid crystals are typically used for opti-
cal displays rather than electro-optic lead lanthanum zirconate titanates (PLZTs).
It is interesting to note that a technology/material which can be used for only one particular
application can succeed in development in a short period. For example, except the display appli-
cation, liquid crystals do not have other applications. From this sense, because there are so many
varieties of applications of ferroelectrics, and the development efforts have become so diverse, that
ferroelectric devices may fail to be developed when competitive materials already exist. Therefore,
we should take a business strategy for the development of ferroelectrics to be focused in the fields
where no other replacement material exists.
In the author’s opinion, the following will be promising areas in the very near future:
1. Electromechanical devices [piezoelectric actuators, ultrasonic motors (USMs)]

2. Thin film hybrid sensors (pyrosensors, pressure sensors, acceleration sensors)
3. Electro-optic devices (light waveguides, thin film hybrid displays)
Of course, this is not meant to discount the other areas of potential development. However, it is
anticipated that the other fields of application will require a higher investment in time, money, and
expertise and a much longer development period than the areas identified as the most promising.
11.1 MARKET SHARE

11.1.1 MARKET SHARE OF FERROELECTRIC DEVICES
Figure 11.1 shows the actual sales of ferroelectric devices (with respect to device type) for the 1983
and 1991 fiscal years in Japan.1,2 To reflect the world market, the data should be multiplied by a
297

1.14 1.18
(1983) (1991)
1.2
Others
(17%)
1.0 Cylinder
(3%)
Sales amount (unit $1 billion)
Semicondu-
0.8 ctor (8%)
0.35 0.61
(1983) (1991)
0.6 Multilayer
(35%) U.S. device (3%)
High-voltage 0.1 0.3
0.4 device (12%) (1983) (1991)
Plate Acoustic device (21%)

0.2 (37%)
Saw device (28%)
Ceramic filter
(36%) PTC
Capacitor Piezoelectric Thermistor

device
FIGURE 11.1 Actual sales amount of ferroelectric devices (with respect to device type) for 1983 and 1991
fiscal years in Japan.
factor of 1.2–1.3, because Japanese industries produce about 80% of these commercial products,
especially the piezoelectric devices. Note that the chart does not include products for military use.
No noticeable change is expected after 1992 due to the following two reasons:
1. Though the sales amount increases yearly, the sales revenue seems to be saturating because
of the price reduction in ferroelectric components.
2. The growth in the new device areas has been suppressed since 1992 due to the worldwide
economical deceleration.
Figure 11.2 shows the ratio of sales by various Japanese manufacturing companies. Note the three
biggest companies which produce a large portion of the ferroelectric-related products: Murata, TDK,
and Matsushita Electronic Components. These three industries share 2/3 of the total production: 3/4
in capacitors, and more than 90% in thermistors. The share ratio is less than 50% only in piezoelectric
devices. Except for products like inexpensive filters, buzzers, etc., the big-three share portion is not
very high in piezoelectric devices, an area which presents much promise for growth in the future.
11.1.2 MARKET SHARE OF PIEZOELECTRIC ACTUATORS

A totally new application area of ferroelectric devices is only the “piezoelectric actuator.” In this
section, we describe the detailed market share of piezoelectric actuators. Table 11.1 summarizes the
developments of ceramic actuators in the United States, Japan, and Europe.3
The developments in the United States are mainly supported by military-related government
institutions and are mainly focused on
1. Active vibration control by army and NASA, using relatively large actuators (>30 cm)
2. Microelectromechanical systems (MEMS) for medical applications and energy
harvesting

Future of Ferroelectric Devices 299
Others
KCK (12%) Others Murata
(7%) Murata (21%) (30%)
(30%)
Matsushita- TDK (3%)
elec. compo. Toko (5%) Matsushita-
(11%) elec. compo.
Taiyo- Sanyo (5%) (17%)
yuden
TDK NGK-NTK Kyocera
(12%)
(28%) (7%) (9%)
Toshiba
(7%)
(a) (b)
Other
(9%) Others
(17%)
Matsushita- Murata
Murata KCK (5%) (30%)
elec. compo.
(19%) (42%)
Kyocer (5%)
TDK
(30%) Taiyo-yuden
(8%)
TDK
(c) Matsushita- (22%)
elec. compo.
(13%)
(d)
FIGURE 11.2 Market share of ferroelectric devices by Japanese manufacturing companies (1983). (a)
Capacitor $1.14 B. (b) Pizoelectric device $0.35 B. (c) Thermister $0.1 B. (d) Total $1.6 B.
TABLE 11.1
Summary of Piezoelectric Actuator Developments, Comparing the United States,
Japan, and Europe
United States Japan Europe
Target Military-oriented product Mass-consumer product Laboratory-equipment product
Category Vibration suppressor Micromotor positioner Micromotor positioner
Vibration suppressor
Application field Space structure Office equipment Laboratory stage/stepper
Military vehicle Camera Airplane
Precision machine Automobile
Automobile Hydraulic system
Actuator Up-sizing Down-sizing Intermediate size
Size 30 cm 1 cm 10 cm
Major AVX/Kyocera NEC-Tokin EPCOS
Manufacturers Morgan Electroceramics Nihon Ceratec Hitachi Metal Siemens Ceram Tec
Micromechatronics Mitsui Chemical Noliac
AlliedSignal Canon Physik Instrumente
Newscale Tech Seiko Instruments Cedrat
Shinsei-Fukoku
Konica-Minolta
Brother

The markets in the United States are limited to military and defense applications, and it is dif-
ficult to estimate the commercial sales amount, because the developments are accelerated through
the government research funds. Among the current needs of the Navy are smart submarine skins,
hydrophone actuators, propeller noise cancellation devices; the Air Force requires smart aircraft
skins; and the Army requires helicopter rotor twisting, aeroservoelastic control, and cabin noise/seat
vibration cancellation devices. Piezoelectric energy harvesting for the ammunition application was
highly accelerated during the new Middle-East War period, but the research focus may be gradually
replaced by medical applications for the treatment of the wounded soldiers.
The piezoactuators and USMs have been developed largely by private industries in Japan, aiming
primarily at applications involving precision positioners and compact motors with tiny actuators of
less than 1 cm for portable electronic equipment. Research and development (R&D) in Japan are
mostly consumer-oriented without being supported by the government. Exceptional national proj-
ects related to this area include “micromechanisms,” which mainly involves the micromachining-
related actuators but with so-called aerosol deposition method from the economical fabrication
viewpoint.
In Japan, piezoelectric camera shutters (Minolta Camera) and automatic focusing mechanisms
in cameras (Canon), dot-matrix printers (NEC), and part-feeders (Sanki) were commercialized and
mass-produced on the order of tens of thousands of pieces per month. Piezoelectric ink-jet print-
ers (Epson), zoom/focus camera modules for mobile phones with an impact mechanism (Konica-
Minolta), and piezoelectric transformers (PTs) (NEC-Tokin, Nihon Ceratec, etc.) are increasing the
sales amount dramatically. A number of patents have been disclosed particularly by NEC, TOTO
Corporation, Matsushita Electric, Brother Industry, Toyota Motors, Tokin, Hitachi Metal, and
Toshiba. The annual sales of ceramic actuator units, camera-related devices, and USMs in Japan
reached $500 million, $300 million, and $150 million, respectively in 2005.4 The total sales may
become equivalent to those of the capacitor industry. If these are installed in final actuator-related
products, sales are projected to reach $10 billion. Thus, a bright future is anticipated in many fields
of application.
European companies started the piezoelectric actuator production about 15 years after Japan and
the United States, and they seemed to have been searching for a variety of possible applications.
The device sizes at the trial manufacturing stage range were generally around 10 cm. However,
during the Japanese Recession and the U.S. War period, European technologies were catching the
Japanese and American levels. Multilayer (ML) piezoactuators by EPCOS and precision stages by
PI (Germany) and Cedrat (France) are some of the most remarkable products.
11.2 RELIABILITY ISSUES

The potential and range of application for ferroelectric materials have been highlighted in the pre-
vious section. However, there still remain various technological problems to resolve before their
full commercial potential can be realized. Of particular concern are the issues of reliability and
durability. Let us consider the reliability issue in terms of reproducibility and lifetime with respect
to materials, device designs, and drive/control techniques.
11.2.1 MATERIALS IMPROVEMENTS

The reproducibility of the dielectric and ferroelectric characteristics of a material depends strongly
on grain size, porosity, and impurity content. Increasing the grain size enhances the magnitude of
the field-induced polarization and strain, but degrades some of the characteristics such as fracture
toughness. The grain size should be optimized for each application. Hence, fine powders made from
wet chemical processes such as coprecipitation and sol-gel will be required.
Porosity must be eliminated completely from the sintered ceramic, when it is used for electro-optic
devices. On the other hand, porosity does not affect the piezoelectric strain behavior significantly, as

long as it is not more than 6%. The tip deflection of a unimorph made from lead magnesium niobate
(PMN)-based material does not change for porosities less than 8%.5 Doping, donor- or acceptor-type,
produces remarkable changes in piezoelectricity. Since “donor” doping (e.g., Nb5+) provides “soft”
characteristics in lead zirconate titanate (PZT), the sample exhibits larger strains and less hysteresis
when driven under a high electric field (1 kV/mm). On the other hand, “acceptor” doping (e.g., Fe3+)
provides “hard” characteristics, leading to a very small hysteretic loss and a large mechanical quality
factor when driven under a small AC electric field (i.e., USM applications).
For most practical devices, the temperature dependence of the characteristics must be stabilized
using either composite or solid solution techniques. The recent trends are in developing extreme
(high or low) temperature sensors and actuators for automobile engines and cryogenic applications
for laboratory equipment and space structures. Systematic studies on the high electric field and
stress dependences of ferroelectric devices as well as the composition dependence of mechani-
cal strength are also eagerly awaited. Other parameters to be considered include “humidity” and
“air pressure,” which is practically essential for the automobile diesel injection valve application.
Electrode silver migration seems to be accelerated with increasing humidity and air pressure, in
addition to the electric field level.
Investigations on aging are also important. The aging effect arises from two factors: depoling
and mechanical destruction. Creep (a gradual change in induced displacements under an applied
voltage) and zero-point drift (zero-field property changes with a cyclic application of voltage) of the
electromechanical characteristics are caused by the depoling of the ferroelectric ceramic with time
lapse. Another serious property degradation is produced by a very high electric field at elevated
temperature, humidity, and mechanical stress, where atomic/ionic migration occasionally occurs.
This may be related either with microcrack propagation or with Ag electrode migration. Nagata
reported that the change in lifetime of a ferroelectric device, such as a ML piezoelectric actuator,
with temperature T and DC bias voltage E obeys an empirical rule6:
TDC = AE − n exp(WDC /kT ), (11.1)
where
WDC is a sort of activation energy
n is a characteristic parameter
Nagata suggested different activation energy W for DC and AC drive conditions.
11.2.2 RELIABILITY OF DEVICES

The popular silver electrodes have a serious problem of migration under a high electric field and
high humidity in actuator, electro-optic, and memory applications, as indicated in the previous
section. This problem can be overcome with the use of a silver–palladium alloy with expensive
palladium higher than 20 mol% (or also expensive Pt). To overcome two problems (atomic migra-
tion and expensive production cost) simultaneously, we need to introduce base metal electrodes,
such as Cu or Ni. However, base metal requires a sintering temperature as low as 900°C and
reduced sintering atmosphere in order to escape from the oxidation problem. Though barium
titanate-based ceramics that can be sintered at low temperature without reduction have been
developed, and Ni electrode has been widely adopted for ML capacitors, the research on Cu
electrode for PZT-based actuator ceramics is still under development for wide commercialization
(refer to Section 4.2.2).
Delamination of the electrode layer is another reliability problem for ML types as well as for
bimorphs. To enhance adhesion strength, composite electrode materials of a metal and ceramic
powder colloid, ceramic electrodes, and electrode configurations via holes are recommended. To
suppress the internal stress concentration which initiates cracks in the device, several electrode

Actuation
Feedback (2)
Breakdown detection
Piezoelectric sensor
actuator
Feedback (1)
Strain sensor
Control
voltage Signal (1)
Signal (2)
Computer-controlled
power supply
FIGURE 11.3 Intelligent actuator system with both position feedback and breakdown detection feedback
mechanisms.
configurations have been proposed: plate-through type, slit-insert type, and float-electrode-
insert type. The reason why the lifetime is extended with decreasing layer thickness has not yet
been clarified.
Lifetime prediction or health monitoring systems using failure detection techniques are also
important for some devices.7 Figure 11.3 shows such an “intelligent” actuator system with acoustic
emission (AE) monitoring. The actuator is controlled by two feedback mechanisms: position feed-
back, which can compensate the position drift and the hysteresis, and breakdown detection feedback
which can stop the actuator system safely without causing any serious damage to the work (e.g.,
in a lathe machine). AE measurement of a piezoactuator under a cyclic electric field is an effective
means of predicting the lifetime of a device. AE is detected primarily when a crack propagates in
the ceramic actuator at maximum speed. Under normal drive conditions of a 100-layer piezoelectric
actuator, the number of AE was counted and a drastic increase, by three orders of magnitude, was
detected just before complete failure. Note that part of the piezodevice can be utilized as an AE
sensor.
A special internal electrode configuration with a strain gauge configuration has been proposed
to increase the reliability of ML piezoelectric actuators.8 As shown in Figure 11.4, strain gauge
TR6847
High voltage supply
FIGURE 11.4 Strain gauge configuration of the internal electrode for an intelligent actuator.

configured electrode patterns are inserted at every 10 internal layers of a multilayer actuator (MLA).
In an electric field cycle normally applied to the device, the resistance change corresponds to the
transverse piezoelectric strain induced in the device. However, if crack or delamination occurs in
the actuator, an abnormally large resistance change is monitored. Thus, this electrode configuration
can be used for both feedback detectors (1) and (2) shown in Figure 11.3.
11.2.3 DRIVE/CONTROL TECHNIQUES

Ferroelectric devices generally have quick responses. However, because of this quick responsive-
ness, when a sharp pulse or step-like voltage is applied to a device, an output overshoot and/or ring-
ing tend to occur just after the voltage is applied. This occurs even in capacitors and electro-optic
devices, where it is sometimes called “screaming” because of the audible sound it sometimes gener-
ates. It originates from a piezoelectrically or electrostrictively induced mechanical resonance.
Accompanying the vibratory output, the pulse drive of the ferroelectric generates large ten-
sile stresses in the device; sometimes large enough to initiate cracks (remember that the ceramic
is weak for the tensile rather than the compressive stress and that the critical stress level for
the PZT is around 100 MPa). In such cases, a compressive bias stress should be applied to the
device with clamping mechanisms such as a helical spring or a plate spring. Piezoactuator struc-
tures for dot-matrix printers and diesel injection valves adopted the stress-bias mechanisms in
practice.
An increase in device temperature is occasionally observed, particularly when the device is
driven by a high alternating electric field, that is, in high-power piezoelectric applications such as
PTs and USMs. The temperature rise is due to the imbalance between heat generation, basically
caused by dielectric hysteresis loss, and heat dissipation, determined by the device size (i.e., sur-
face area).9 It is necessary to select a suitable duty ratio for the drive power so as not to generate a
temperature rise higher than 30°C. The standard high-power piezoelectric measurement method
suggests a maximum temperature rise of 20°C for determining the maximum vibration velocity
(a little conservative standpoint).
As far as high-power ultrasonic transducers and motors are concerned, operation in the anti-
resonance mode has been proposed.10 USMs have conventionally been operated in the reso-
nance mode, at the so-called resonance frequency. However, the mechanical resonance state
at the “antiresonance” frequency reveals much higher Q M and lower heat generation than those
observed for the “resonance” condition, when the same vibration amplitude (velocity) is excited.
Moreover, the operation at “antiresonance,” where admittance is very low, requires low current
and high voltage for driving, in contrast to high current and low voltage for the resonance opera-
tion. This means that a conventional inexpensive power supply may be utilized for driving the
ultrasonic device.
11.2.4 SAFETY SYSTEMS

Future R&D should focus on ecologically superior systems (i.e., those fit for humans) as well as
technologically advanced ones (i.e., just for the performance). PZT-based ceramics were the key
players highlighted in this textbook. However, the development of lead-free, nontoxic ferroelectric
ceramics will certainly be even more important in the twenty-first century, when we have some polit-
ical regulation on materials containing “lead.” Pb-free ceramics, such as BaTiO3- and K(Ta,Nb)O3-
based materials, are being studied vigorously at present, particularly in the fields of medical and
automobile applications (which may request “disposable” characteristic).
Safety systems, which can both monitor the fatigue or symptom of failure of materials/devices
and stop the equipment safely without causing serious problems, are also desired. A strain-gauge
internal electrode configuration for ML piezoelectric actuators is a good example of a future safety
system.

11.3 DEVELOPMENT OF BESTSELLING DEVICES

11.3.1 THREE CREATIVITIES
In mid-1980, Akio Morita, former president of SONY Corporation, responded to criticism from a
journalist concerning the lack of creativity on the part of Japanese researchers by saying “Japanese
researchers are good at chasing and imitating the original idea for commercialization, but they in
general lack creativity.”11 Mr. Morita suggested that there should be three types of creativity with
respect to R&D at SONY: “The U.S. people are focusing only on technological creativity. But the
people must understand there are two more creativities; product planning creativity and marketing
creativity, which are equally important for commercial success.”
Matsushita Panasonic’s famous color TV technology (black stripe for creating better color reso-
lution) was invented by Philips. Philips could not commercialize it. Matsushita, on the other hand,
succeeded after an intensive three-year development effort. You can decide which company is more
“creative” in science and technology, however only Matsushita profited from this TV development.
In this last section, the author presents the reader his personal business philosophy on how to
develop bestselling devices, from his 35 years of academic experiences and 20 years of company
executive experiences. Table 11.2 summarizes initially these three important types of creativity
needed when developing an R&D strategy; each will be described in further detail in the fol-
lowing sections. We will consider these creativities in the sequence shown in Table 11.2. A new
technology or an idea for a new device is the first step for a “high-tech entrepreneur.” Next comes
the product design and commercialization plan for the first prototype. Finally, marketing promo-
tion is important for making the product a bestseller. This development style is called seed-push.
Note that this sequence is sometimes reversed in a big firm; first market research finds a product
need, which is fed back to technology development, and finally commercialized. This is called
need-pull.
11.3.2 TECHNOLOGICAL CREATIVITY

There are two different approaches in exercising technology creativity: (1) Find a new functional
effect or material and (2) Achieve a high performance or figure of merit (FOM). These are typically
called “research” and “development,” respectively. A new idea arising from “research” will create a
seed-push market, while development is initiated by a need-pull force.
11.3.2.1 Discovery of a New Function or Material

Serendipity is often an important factor in discovering a new function in a material or a new phe-
nomenon. Benjamin Franklin the famous scientist and founding father of the United States, dis-
covered that lightning is an electrical phenomenon.12 Franklin’s experiment was done in a thunder
TABLE 11.2
Three Types of Creativity in R&D
1. Technological creativity New functions
High performance
2. Product planning creativity Specification (sensitivity,
size, power)
Design
3. Marketing creativity Price
Advertisement

FIGURE 11.5 Flying kite experiment in a thunderstorm by Ben Franklin (1752). (From http://www.
ushistory.org/franklin/essays/hoffman.htm. With permission.)
storm. The lightning hit the kite, and he collected electric charge, as illustrated in Figure 11.5. Two
other scientists conducting similar experiments one month before and after Franklin were both
electrocuted. Franklin was one lucky person. Franklin became President of Pennsylvania (equiva-
lent to Governor today) and the state organization for financially supporting incubator companies in
Pennsylvania is named “Ben Franklin Technology Partners.”
“Ivory” soap by P&G was created also by serendipity. William Procter and James Gamble
started a candle shop, but because of Thomas Edison’s light bulbs, their business declined. In 1879,
an employee in their Cincinnati candle factory forgot to turn off a machine when he went to lunch.
On returning, he found a frothing mass of lather filled with air bubbles. He almost threw the stuff
away, but instead decided to make it into soap. They sold the soap as a “floating” soap with a lot of
bubbles. Why was floating soap such a hot item back then? Because clothes were washed in ponds
and rivers at that time. A dropped bar of soap would sink and often be unrecoverable. Floating soap
had a convenience factor.
PVDF, a piezoelectric polymer was discovered accidentally in early 1970s by Dr. Kawai, and a
high-temperature superconducting ceramic was discovered by Dr. Bednortz and Dr. Muller in 1980.
Both are good examples of serendipity in the related field of ferroelectric materials.
A traditional Japanese proverb tells us every researcher has three lucky chances in his/her life
to discover new things. However, most people do not even recognize these chances and lose them.
Only people ready to accept these chances can really find new phenomenon. A Japanese company
executive mentioned that a person who develops “one” widely commercialized product has the
chance to become a general manager; a person who develops “two” products for the company is
guaranteed to be a vice president; and a person who contributes more than “three” can become
president. You can see how difficult it is to develop actual bestselling products.
Suppose that you are an engineering student, or a young researcher in a manufacturing company,
are you confident to continue your career path? Or, are you feeling a slump in your R&D activity,
and are you more likely to be comfortable in a management position? Example Problem 11.1 will
help in assessing your creative aptitude.


Figure 11.6 is a test picture with text randomly cited from an academic journal.
1. First, familiarize yourself with the contents and picture of the academic article below
(Figure 11.6) for 60 s. Do not peek at the following page, which includes the Questions and
Answers. Remember that this is your own personality check before reading this textbook
further.
After checking your wristwatch, you may start NOW. Move to the next page after 60 s.
Kenji Uchino is an enthusiastic explainer

of the mysteries of miniaturization who
manages to not only explain complex
subjects, but to get his listeners caught up
in the excitement of discovery. In the
classroom, he introduces students to the
joys of ferroelectricity. In industry, he has
collaborated in the development of dozens
of specialized patented devices.
15
FIGURE 11.6 Test picture. (Note that this article was randomly cited from an academic journal.)

2. Second, answer True or False for the following sentences without seeing the previous
page:
a. His name is Ohuchi.
b. His article is printed on p. 15 of an academic journal.
c. He has moustache.
d. He wears a dotted-design tie.
3. Third, score 1 point for each correct answer according to the solution box lower on the
page.
Comments
Your Score Aptitude

Really recognized 4 You can be a good engineer
Some guesses 2∼3 You fit to a manager/sales engineer
No idea 0 Abandon your dream to be an engineer
People with engineering aptitude generally remember the written content first. If you have
failed to answer questions (a) and (b) correctly, then your aptitude in this direction is lacking.
Recognizing the “moustache” is also expected of engineering types, because it directly belongs
to the person. However, remembering the tie is relatively rare. You can see it only when you
try to. This attention to detail is one of the most important aptitudes to cultivate for innovative
creativity.
This can be used to think up unconventional questions for a job interviewee. The following
are two questions I have used:
a. You climbed up a staircase a couple of minutes ago. How many stairs did you climb?
b. You must have seen the pedestrian traffic signal just before entering the company entrance.
Do you remember an illustration of a walking man lit up in blue? Is he walking toward the
left or toward the right?
For the second question, most of the interviewees recognize the illustration, but the answers
differ remarkably. When the answer is “I don’t remember,” we usually find the candidate unsuit-
able. Even when the answer is correct, “left,” it may be a guess with a probability of 50%. This
candidate may be hired for a management position. Only when the correct answer arises from
a confident memory, will we hire the candidate as a professional engineer. I occasionally use a
similar test for hiring suitable people for the university research center and/or the company. Once
you become a corporate officer or a manager of your new venture who is responsible for hiring
research and engineering employees, you may need to create similar test questions.
If you have missed your “three chances,” what should you do? Quit being a researcher? The fol-
lowing example is dedicated to an unlucky reader, who, like the author, missed those lucky chances.
We can still research using a more systematic approach, for example, by using our intuition: Making
use of (a) secondary effects and (b) scientific analogy.
a. Secondary effect
Every phenomenon has primary and secondary effects, which are sometimes recognized
as linear and quadratic phenomena, respectively. In electro-optic devices, the Pockels and
Kerr effects correspond to the primary and secondary effects. In actuator materials, these
correspond to the piezoelectric and electrostrictive effects.
When the author started actuator research in the mid-1970s, precise “displacement trans-
ducers” (we initially used this terminology) were required in the Space Shuttle program,
Solution: (1) F, (2) T, (3) T, (4) F

X
X
E
E
(a) (b)
FIGURE 11.7 (a) Primary effect (piezoelectric effect) and (b) secondary effect (electrostrictive effect).
particularly for “deformable mirrors,” for controlling the optical pathlengths over several
wavelengths (0.1 μm). Conventional piezoelectric PZT ceramics were plagued by hyster-
esis and aging effects under large electric fields; this was a serious problem for an optical
positioner. Electrostriction, which is the secondary electromechanical coupling observed
in centrosymmetric crystals, is not affected by hysteresis or aging (see Figure 11.7). Its
response should be much faster than the time required for domain reorientation in piezo-
electrics/ferroelectrics. In addition, electric poling is not required.
However, at that time, most of the people believed that the secondary effect would be
minor, and could not provide a larger contribution than the primary effect. Of course, this
may be true in most cases, but, the author’s group actually discovered that relaxor ferro-
electrics, such as the PMN-based solid solutions exhibit enormous electrostriction.
b. Scientific analogy
Probably most readers are familiar with shape memory alloys, which can revert rather
quickly back to their initial shape when subjected to the heat of a cigarette lighter or a hair
dryer. The basic principle is a “stress or temperature-induced” phase transformation from
the austenite to martensite phase. The author tried to consider an analogous case among
the ferroelectrics (Figure 11.8). Yes, we have an “electric-field-induced” phase transition
from an antiferroelectric to ferroelectric phase. This type of phase transition should be
much quicker in response and more energy efficient theoretically. After this speculation,
we started to investigate lead zirconate-based antiferroelectrics intensively, and discovered
the “shape memory effect” in ceramic actuator materials.
11.3.2.2 Performance Improvement

Starting with material functionality, Table 11.3 lists the various effects relating input (electric field,
magnetic field, stress, heat, and light) with output (charge/current, magnetization, strain, temperature,
and light). Conducting and elastic materials, which generate current and strain outputs, respectively, for
input voltage or stress are well-known phenomena. They are sometimes called trivial materials. On the
other hand, pyroelectric and piezoelectric materials, which unexpectedly generate an electric field with
the input of heat and stress, respectively, are called smart materials. These off-diagonal couplings have
corresponding converse effects, the electrocaloric and converse-piezoelectric effects. Both “sensing”
and “actuating” functions can be realized in the same materials. “Intelligent” materials must possess a
“drive/control” or “processing” function which is adaptive to changes in environmental conditions, in
addition to actuator and sensing functions. Ferroelectric materials exhibit most of these effects with the
exception of magnetic phenomena. Thus, ferroelectrics are said to be very “smart” materials.
The concept of composite effects, learned in Chapter 10 is very useful, particularly for systemati-
cally improving the properties and FOM.
a. Combination effect
In certain cases, the averaged value of the output, Y*, of a composite does exceed Y1 and
Y2. This enhanced output refers to an effect Y/Z which depends on two parameters Y and Z.
Suppose that Y and Z follow convex- and concave-type sum effects, respectively, as illus-
trated in Figure 11.9, the combination value Y/Z will exhibit a maximum at an intermediate
ratio of phases. This is called a combination effect. Certain piezoelectric ceramic/polymer

Parent
phase
Cool Heat
(Initial state) (Deformed state)

Stress
applied
and removed
Martensitic Deformed
phase martensitic
phase
(a) Shape memory alloy
Paraelectric
phase
Cool Heat
Inverse
field
applied
and removed
(Initial state) (Deformed state)
Field
applied
Antiferroelectric and removed
Ferroelectric
phase
phase
(b) Shape memory ceramic
FIGURE 11.8 Phase transition analogy between (a) shape memory alloy and (b) shape memory ceramic.
composites exhibit a combination property of g (the piezoelectric voltage constant) which

is provided by d/ε0ε (d: piezoelectric strain constant, and ε: relative permittivity).
b. Product effects
When Phase 1 exhibits an output Y with an input X, and Phase 2 exhibits an output Z with
an input Y, we can expect for the composite an output Z with an input X. A completely new
function is created for the composite structure, called a product effect.

TABLE 11.3
Various Effects in Materials
Materal
Input Output
device
Output Charge Magnet- Strain Temperature Light

Input current ization
Permittivity Elect.-mag. Converse Elec. caloric Elec.-optic

Elec. field effect
Conductivity piezoeffect effect effect
Mag.-elect. Permeability Magneto- Mag. caloric Mag.-optic
Mag. field effect effect
effect striction
Piezoelectric Piezomag. Elastic ___ Photoelastic
Stress effect constant
effect effect
Pyroelectric ___ Thermal Specific ___
Heat effect expansion heat
Photovoltaic ___ ___ Refractive
Light effect Photostriction
index
Diagonal coupling Sensor

Off-diagonal coupling = Smart material Actuator
Phase 1 : X Y1/Z1
X (Y/Z)*
Phase 2 : X Y2/Z2
Improvement
Y1
Y2
Phase 1 Phase 2
Y1/Z1 Y2/Z2
Z1
Phase 1 Phase 2
Z2
Phase 1 Phase 2
FIGURE 11.9 Basic concept of the performance improvement in a composite via a combination effect.
The author introduces his Functionality Matrix concept here. If one material has “piezo-
magnetic effect” and its converse effect “magnetostrictive effect,” the functionality matrix
of this material can be expressed by
⎛0 0 0 0 0⎞
⎜0 0 Magnetostriction 0 0⎟
⎜ ⎟
⎜0 Piezomagnetic effect 0 0 0⎟ .
⎜0 0 0 0 0⎟
⎜ ⎟
⎜⎝ 0 0 0 0 0⎟⎠

On the other hand, a piezoelectric has a functionality matrix of the following form:
⎛ 0 0 Converse piezoeffect 0 0⎞
⎜ 0 0 0 0 0⎟
⎜ ⎟
⎜ Piezoelectric effect 0 0 0 0⎟ .
⎜ 0 0 0 0 0⎟
⎜ ⎟
⎝ 0 0 0 0 0⎠
When the magnetic field is input first, the expected phenomenon is expressed by the matrix product:
⎛0 0 0 0 0⎞
⎜0 0 Magnetostriction 0 0⎟⎟
⎜
⎜0 Piezomagnetic effect 0 0 0⎟
⎜ ⎟
⎜0 0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0 0⎠
⎛ 0 0 Converse piezoeffect 0 0⎞ ⎛ 0 0 0 0 0⎞
⎜ 0 0 0 0 0⎟⎟ ⎜⎜ Mag.-elect.effect 0 0 0 0⎟⎟
⎜
⊗ ⎜ Piezoelectric effect 0 0 0 0⎟ = ⎜ 0 0 0 0 0⎟ .
⎜ ⎟ ⎜ ⎟
⎜ 0 0 0 0 0⎟ ⎜ 0 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⎝ 0 0 0 0 0⎠ ⎝ 0 0 0 0 0⎠
If we start from the electric field input first, the expected phenomenon will be
⎛ 0 0 Converse piezoeffect 0 0⎞
⎜ 0 0 0 0 0⎟
⎜ ⎟
⎜ Piezoelectric effect 0 0 0 0⎟
⎜ 0 0 0 0 0⎟
⎜ ⎟
⎝ 0 0 0 0 0⎠
⎛0 0 0 0 0⎞ ⎛0 Elect.-mag. effect 0 0 0⎞
⎜0 0 Magnetostriction 0 0⎟ ⎜0 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⊗⎜0 Piezomagnetic effect 0 0 0⎟ = ⎜0 0 0 0 0⎟ .
⎜0 0 0 0 0⎟ ⎜0 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 0 0⎠ ⎝0 0 0 0 0⎠
Note that the resulting product matrixes include only one component each; magnetoelectric effect
or electromagnetic (EM) effect component.
Philips developed a magnetoelectric material based on this concept.13 This material is composed
of magnetostrictive CoFe2O4 and piezoelectric BaTiO3 mixed and sintered together. Figure 11.10a
shows a micrograph of a transverse section of a unidirectionally solidified rod of materials with
an excess of TiO2 (1.5 wt.%). Four finned spinel dendrites CoFe2O4 are observed in cells (×100).
Figure 11.10b shows the magnetic field dependence of the magnetoelectric effect in an arbitrary unit
measured at room temperature. When a magnetic field is applied on this composite, cobalt ferrite
generates magnetostriction, which is transferred to barium titanate as stress, finally leading to the
generation of a charge/voltage via the piezoelectric effect in BaTiO3.

ΔE
ΔH
Hmax Hdc
(b)
FIGURE 11.10 (a) Micrograph of a transverse section of a unidirectionally solidified rod of mixture of mag-
netostrictive CoFe2O4 and piezoelectric BaTiO3 with an excess of TiO2. (b) Magnetic field dependence of the
magnetoelectric effect in a CoFe2O4:BaTiO3 composite (at room temperature).
The author’s photostrictive materials were also discovered along a similar line of reasoning;
functionality matrices of photovoltatic and piezoelectric effects. The following is an anecdote from
The R&D Innovator.14
I’ve made a breakthrough that could lead to photophones–devices without electrical connections that
convert light energy directly into sound. Perhaps this discovery will help commercialize optical tele-
phone networks. It also could allow robots to respond directly to light; again, without a need for wire
connectors.
Where did I come up with the idea for this light conversion? Not with the sunlight shining through my
office window, and not outside feeling the warmth of the sun, but in a dimly lit Karaoke bar.
I’ve been working on ceramic actuators–a kind of transducer that converts electrical energy to mechan-
ical energy–at the Tokyo Institute of Technology when the trigger for “the light-controlled actuator”
was initiated. In 1980, one of my friends, a precision-machine expert, and I were drinking together
at a Karaoke bar, where many Japanese go to enjoy drinks and our own singing. We call this activity
our “after-5-o’clock meeting.” My friend studied micro-mechanisms such as millimeter-size walking
robots. He explained that, as electrically controlled walking mechanisms become very small (on the
order of a millimeter), they don’t walk smoothly because the frictional force drops drastically and the
weight of the electric lead becomes more significant.
After a few drinks, it becomes easier to play “what if?” games. That‘s when he asked, “What if you, an
expert on actuators, could produce a remote-controlled actuator? One that would bypass the electrical
lead?” To many people, “remote control” equals control by radio waves, light waves, or sound. Light-
controlled actuators require that light energy be transduced twice: first from light energy to electrical
energy, and second from electrical energy to mechanical energy. These are “photovoltaic” and “piezo-
electric” effects.
A solar cell is a well-known photovoltaic device, but it doesn’t generate sufficient voltage to drive a
piezoelectric device. So my friend’s actuator needed another way to achieve a photovoltaic effect. Along
with the drinking and singing, we enjoyed these intellectual challenges. I must have had a bit too much
that night since I promised I’d make such a machine for him. But I had no idea how to do it!
While my work is applied research, I usually come home from scientific meetings about basic research
with all kinds of ideas. At one of these meetings, about six months after my promise, a Russian physicist
reported that a single crystal of lithium niobate produced a high electomotive force (10 kV/mm) under
purple light. His talk got me excited. Could this material make the power supply for the piezoelectric
actuator? Could it directly produce a mechanical force under purple light?
I returned to the lab and placed a small lithium niobate plate onto a plate of piezoelectric lead zircon-
ate titanate. Then I turned on the purple light and watched for the piezoelectric effect (mechanical

deformation). But it was too slow, taking an hour for the voltage to get high enough to make a
discernable shape change.
Then the idea hit me: what about making a single material that could be used for the sensor and the
actuator? Could I place the photovoltaic and piezoelectric effects in a single asymmetric crystal? After
lots of trial and error, I came up with a tungstate-doped material made of lead lanthanum zirconate
titanate (PLZT) that responded well to purple light. It has a large piezoelectric effect and has properties
that would make it relatively easy to fabricate.
To make a device out of this material, I pasted two PLZT plates back to back, but placed them in oppo-
site polarization, then connected the edges. I shined a purple light to one side, which generated a pho-
tovoltaic voltage of 7 kV across the length. This caused the PLZT plate on that side to expand by nearly
0.1% of its length, while the plate on the other (unlit) side contracted due to the piezoelectric effect
through the photovoltage. The whole device bent away from the light. For this 20 mm long, 0.4 mm
thick bi-plate, the displacement at the edge was 150 μm, and the response speed was 1 second. This fast
and significant response was pretty exciting.
Remembering the promise to my friend, I fabricated a simple “light-driven micro walking machine,”
with two bi-plate legs attached to a plastic board, as shown in Figure 11.11. When light alternately irra-
diated each leg, the legs bent one at a time, and the machine moved like an inchworm. It moved without
electric leads or circuits! That was in 1987, seven years after my promise.
I got busy with my “toy”; but not too busy to attend “after-5-o’clock meetings” in Tokyo’s night-club
area. In 1989, at my favorite Karaoke bar, I was talking about my device to another friend who worked
for a telephone company. He wanted to know if the material could make a photo-acoustic device–
perhaps as a solution to a major barrier in optical-fiber communication. The technology to transmit
voice data–a phone call–at the speed of light through lasers and fiber optics has been advancing rapidly.
But the end of the line–the ear speaker–limits the technology, since optical phone signals must be con-
verted from light energy to mechanical movement via electrical energy.
I thought my material could convert light flashes directly into sound. I chopped two light beams to make
a 180-degree phase difference, and applied each beam to one side of the bi-plate. The resonance point,
monitored by the tip displacement, was 75 Hz, just at the edge of the audible range for people! We’re
now working to fabricate real photo-speakers (I call them “photophones”), and have ideas that may
FIGURE 11.11 Photo-driven walking machine.

increase the vibration frequency several-fold to reproduce human speech correctly. Photophones could
provide a breakthrough in optical communication.
Well, what’s my message for you, dear reader? To find a noisy Karaoke bar? Perhaps that’s not neces-
sary; but what is necessary is listening to others outside your particular research area: for instance,
basic researchers or people with specific, applied objectives.
The above anecdote indicates another important issue; discovery was motivated by strong cus-
tomer demand. This is a good example of a “Need-Pull” development. The situation of discover-
ing “monomorphs” (semiconductive piezoelectric bending actuators) is also similar to the above.15
When attending a basic conference of the Physical Society of Japan, I learned about a surface
layer generated on a ferroelectric single crystal due to the formation of a Schottky barrier. It was
not difficult to replace some of the technical terminologies with our words. First, polycrystalline
piezoelectric samples were used, with reduction processes to expand the Schottky barrier thick-
ness. We succeeded in developing a monolithic bending actuator. The “Rainbow” structure, further
developed by Aura Ceramics, is one of the monomorph modifications.
11.3.3 PRODUCT PLANNING CREATIVITY

11.3.3.1 Seeds and Needs
I usually suggest product planning divisions reexamine 10-year-old research. If the social needs
still exist, there will probably be a good business opportunity because the related patents have prob-
ably expired or will soon expire. More importantly, find the reasons for lack of success and judge
your company’s capability to overcome them.
Examples can be found in 2D displays and piezoelectric transformers. In collaboration with
Fujitsu General, Japan, the author’s group developed a 2D PLZT display using the newly developed
nanopowder technology and tape casting. The idea of the 2D display existed from 1970s, but there
has not been the fabrication technology by then. Our effort for preparing sophisticated PLZT fine
powders made this device realizable. We always need to watch carefully for possible supporting
technological development.
Figure 11.12 shows the number of yearly patent disclosure relating to PTs from 1972 till 1999.
There are clear two peaks: 1972 and in 1998. During the 25–30 year gap there was almost no devel-
opment. Historically, PTs were commercialized for the first time in the beginning of the 1970s to
supply high voltage in color TVs. However, this application disappeared in less than a year, due to
cracking which destroyed the devices. The PZT ceramic mechanical quality was so poor to endure
1972
1976
1980
Year
1984
1988
1992
1996
1999
0 5 10 15 20 25
Numbers
FIGURE 11.12 Number of yearly patent disclosure relating to PTs from 1972 till 1999.

for the commercial product at that time. The second commercialization peak occurred due to three
key factors: strong social demand for a laptop computer backlight screen, matured powder technol-
ogy to provide mechanically strong piezoceramics, and advanced design technology such as finite
element method software to simulate electromechanical vibrations. The author is proud to say that
his 35 years of contribution is mostly dedicated to enhance the piezoceramic reliabilities. Note also
that once one device development fails, a similar development may not restart in 25–30 years, which
is known as the R&D cycle and will be discussed again in the next section.
Tracking future technologies is also important in finding “seeds” for new products. Battelle
reports regularly on future technologies. Their 1995 top 10 predictions for 2005 are16
1. Human genome mapping. Genetic-based personal identification and diagnostics will lead
to preventive treatments of disease and cures for specific cancers.
2. Super materials. Computer-based design and manufacturing of new materials at the molec-
ular level will mean new, high-performance materials for use in transportation, computers,
energy, and communications.
3. Compact, long-lasting, highly portable energy sources, including fuel cells and batteries,
will power electronic devices of the future, such as portable personal computers.
4. Digital, high-definition TV. A major breakthrough for American television
manufacturers—and a major source of revenue—that will lead to better advanced com-
puter modeling and imaging.
5. Electronics miniaturization for personal use. Interactive, wireless data centers in a pocket-
size unit will provide users with a fax machine, telephone, and computer that contains a
hard drive capable of storing all the volumes found in their local library.
6. Cost-effective “smart systems” will integrate power, sensors, and controls. These systems
will eventually control the manufacturing process from beginning to end.
7. Antiaging products—that rely on genetic information to slow the aging process—will
include antiaging creams that really work.
8. Medical treatments that will use highly accurate sensors to locate problems, and drug-
delivery systems that will precisely target parts of the body, such as chemotherapy targeted
specifically to cancer cells to reduce the side effects of nausea and hair loss.
9. Hybrid-fuel vehicles. Smart vehicles, equipped to operate on a variety of fuels, will be able
to select the most appropriate one based on driving conditions.
10. “Edutainment.” Educational games and computerized simulations will meet the sophisti-
cated tastes of computer-literate students.
As of today (2009), the reader can understand that Battelle’s prediction hit rate is very high (90%).
Thus, you can use Battelle’s predictions as a surrogate if you do not have the resources for doing
this yourself.
In general, microactuators will be required for medical diagnostic applications such as blood test
kits and surgical catheters. Though silicon MEMS are developing rapidly, pure electrostatic forces
are, generally, too weak to move something mechanically with reasonable efficiency. These practi-
cal actuators need to generate the mechanical energy to fight with human finger force. Piezoelectric
thin films compatible with silicon technology will be much more useful for MEMS because the
energy density can be enhanced by a factor of 100. An ultrasonic rotary motor as tiny as 2 mm in
diameter, fabricated on a silicon membrane, is a good example (see Figure 11.13).17 Even this proto-
type motor can generate a torque three to four orders of magnitude higher than an equivalent size
silicon motor.
As the size of miniature robots and actuators decrease, the weight of the electric lead wire con-
necting the power supply becomes significant, and remote control will definitely be required for
submillimeter devices. The photo-driven actuator described in the previous section is a promising
candidate for microrobots in the future.

Glass lens rotor

Gold top electrode
Ti/Pt bottom
PZT thin film electrode
Silicon
Silicon nitride
FIGURE 11.13 Ultrasonic rotary motor as tiny as 2 mm in diameter fabricated on a silicon membrane. (From
Flyn, A.M. et al., J. Microelectromech., Syst., 1, 44, 1992.)
11.3.3.2 Innovation Obstacle in Technology Management

Innovation obstacles are important issues for technology management. We have observed a pattern:
once development has failed, a minimum of 25 ∼ 30 years is required to redevelop the same/similar
device. The reason is once a young researcher fails in development, the experience stops him from
engaging in similar device development or allowing others in his organization from doing the same
after he becomes a manager. It takes a generation, 25 years, for the technology to be resurrected,
until these “fossil” managers disappear (or retire).
Two issues to remember:
1. If you are a manager, try not to suppress redevelopment of previously failed devices.
2. If you are a young researcher, reexamine 10-year-old research (a half of a generation
period). If your boss is a “fossil” type, you had better spin off from that institute, and start
your own company.
11.3.3.3 Development Pace

A suitable R&D pace introduces new concepts and products neither too early nor too late. Three
years for commercialization is a good target for the ferroelectric devices. The Ford Motor Company
changed their development pace from 5 to 3 years several years ago, and commercialized the
“Taurus” successfully.
11.3.3.4 Specifications
Some engineers believe that lowering the drive voltage of a piezoelectric actuator is essential.
However, this is not really true for portable equipment if one considers the available battery volt-
ages. Does the reader know the available battery voltages and voltage suppliers? The answers are
1.5, 3, 6, 12 (automobile applications), 24, and 250 V. For example, when I collaborated with COPAL
to develop piezoelectric camera shutters using a bimorph structure, we initially used conventional
bimorphs driven at around 100 V (see Figure 11.14). But, when we tried to commercialize it, we
recognized that we needed an additional 100 V power supply in each camera, which would cost sev-
eral dollars. Instead we changed the bimorph design, by thickening the piezoelectric ceramic layer,
so that it could be driven by 250 V (this voltage is generated in a camera by a cheap power supply
conventionally used for a stroboscopic lamp). Remember that even one resistor cannot be employed
additionally for a cheap camera application.
Product development needs to collect the necessary information on the specifications:
• Sensitivity
• Size
• Lifetime
• Available power supply

Bimorph support
Wing B
Wing
pivot
Bimorph
Shutter opening
Lever
Wing A
Closed state Open state
FIGURE 11.14 Piezoelectric bimorph camera shutter by COPAL, where the PZT layer was intentionally
thickened to adapt to a 250 V voltage supply.
If we extend the above consideration, a device for automobiles should be driven under 12 V (car
battery). However, actual piezoelectric actuators for diesel injection valves applications are driven
with 160 V for 80 μm PZT layer thickness (Figure 11.15a). The key restriction not to allow the drive
voltage below 12 V per 10 μm PZT layer is not from technological difficulties, but merely from the
cost minimization. Figure 11.15b shows prices for both the piezoelectric stack actuator and its drive/
control circuit, plotted as a function of drive voltage. While decreasing the drive voltage, the PZT
layer thickness should be decreased dramatically, leading to an increase in the Ag/Pd electrode
cost and manufacturing costs increase exponentially. On the other hand, the driving circuit cost
increases with the required voltage. Accordingly, the total system cost has a minimum around 160 V
and a layer thickness around 80 μm. Note that even 10 μm layer actuators are not difficult to manu-
facture today. The present specification for thickness is merely determined by economics!
Piezo injector
High energy (high
current capacitance)
Injector
Thin layer ML
Discrete and EDU cost
head
Piezo stack price
Injector Low energy (low

body current capacitance)
Nozzle Thick layer
10 V 100 V 1000 V
(a) (b)
FIGURE 11.15 (a) Piezoelectric diesel injection valve by Siemens, (b) cost evaluation for the piezostack and
drive circuit as a function of drive voltage. (From Siemens News Release; Fujii, A., Proc. JTTAS Meeting
Dec. 2, Tokyo, 2005. With permission.)

11.3.3.5 Product Design Philosophy

When performance is similar, sales depend strongly on design, color, etc. The design must fit social
trends as discussed in detail in Section 11.3.4. In this new era, Y. Hirashima suggests the keywords
beautiful, amusing, tasteful, and creative for new products.18 General “Managerial Economics”
textbook19 suggests that product demand is determined by six variables: price, income of consum-
ers, prices of related products, expected future price of product, number of consumers, and taste
patterns of consumers. The last factor has been emphasized in the last 10 years.
Table 11.4 summarizes the differences in development concepts between the United States and
Japan. Samsonite had a TV commercial about its suitcases 10 years ago. In a 3 min broadcast,
a suitcase is thrown away from a 10-storey building without damage. This ad provides a show how
highly reliable and tough the suitcase is. I appreciate this typical American attitude, but concerned
that they are not with the contents in the suitcase, such as glass bottles. A similar concept can be
found in a recent TV ad by Maytag. Their washing machine ad seems ridiculous. Lots of tennis balls
hit the surface of the washing machine, which may show its mechanical toughness. Actually the
purpose of a washing machine is how quickly, silently and efficiently it cleans clothes. The ad did
not mention anything about its basic performance, merely showed its mechanical toughness.
The promotion of Nintendo’s Game Boy is based on very different concepts; newness and timing.
The game might not have sold well due simply to misjudgment in timing. Moreover, when Nintendo
developed a “family computer,” they seemed to try to weaken the connector between the software
board and the hardware chassis. It seemed to become damaged after about 100 connections, which
corresponds to 3–6 month normal usage for a child. When the toy breaks and a child complains to
his/her parent, the response (particularly Japanese) may be, “You play computer games too much!
It is about time you returned to your studies!” In this scenario, few parents would bother to com-
plain to Nintendo. Of course, the child wants it fixed, once he knows how much fun it is. He may
even spend his own money to purchase another game machine; if this is true, the author admires
Nintendo’s strategy. Recent game equipment has dramatically improved, so this is no longer true,
but it occurred in the older models.
This basic consumer attitude difference explains why e-mail on a mobile phone system became
popular in Japan 10 years ahead of the United States. Camera-phone advertisements are finally
broadcasted in the United States in recent years. Americans like Walmart because of “lowest” price,
while Walmart cannot invade Japan because of low newness. They may purchase old products from
manufacturers at the lowest cost.
11.3.3.6 Smart Systems

“Intelligent” or “smart” materials, structures, and systems are often used today. The bottom line
of “smartness” is to possess both “sensing” and “actuating” functions. I offered my opinion on this
issue in Chapter 1.
When a new “sensing” function is required, most researchers try adding an additional component,
leading to a more complex system, which is likely to be more bulky and expensive (I call it “Spaghetti
TABLE 11.4
Difference in Development Concepts between
the United States and Japan
United States Japan
Best device Good device
For military For civil consumers
With top technology With improved technology
At expensive cost At cheap cost
(Reliability) (Newness, timing)

Multilayer piezo actuator
E
Higher function
C
Add components
Actuator
Preload
Material/system (Ultrasonic motor)
Duralumin
Reduce components (Combine sensors and
actuators)
Ra
il
Smaller size, lower cost
Multilayer
Actuator
piezoelectric
actuator
Log
Rail
FIGURE 11.16 Development concept for smart systems, using USMs as an example.
Syndrome”). My group is contributing enormously to adding new functions to conventional materials

and structures, while reducing the number of components in a system, while aiming for miniaturiza-
tion and lower cost. The photostrictive actuator is a very good example of an “intelligent” material. It
“senses” light illumination and generates a voltage/current proportional to the light intensity. Then it
produces strains according to this “control” voltage, leading to the final mechanical “actuation.”
The design/development concept for smart systems is illustrated in Figure 11.16, using USMs.
Starting from a propagating-wave type motor with two piezoactuators and two power supplies, a
group of Philips moved to a more complex motor with four piezoactuators in order to seek better
motor performance. However, our group took the opposite approach, simplification, and developed
a standing-wave type with a single actuator element in order to make it smaller and less expensive,
and it was successfully commercialized for x–y stages. It is your choice to seek merely performance
(for mostly Military or Space applications) or to seek the optimization of (performance/cost) (for
general consumer applications).
11.3.4 MARKETING CREATIVITY

“Discipline of Market Leaders”20 authored by Treacy and Wiersema is a very informative guide to
understanding marketing creativity. They have factored it into three basic steps:
a. Choose your customers

b. Narrow your focus
c. Dominate your market
We will discuss in detail these steps.

11.3.4.1 Choose Your Customers

a. Domestic or foreign?
Let us start by solving the following practical problems which arose during my consulting.

a. Japanese cars are popular in the United States, but U.S. cars are not popular in Japan,
why?
b. TOTO “Washlet” (a personal hygiene system, for cleaning one’s private parts, in the toilet
facility) is a big hit in Japan, but not in the United States, although TOTO Co. tried to sell
it in the United States, why?
Solution
a. There is a traffic system difference between the two countries: left-hand side in Japan
and right-hand side in the United States. While the Japanese auto manufacturers tried
to make left-hand side steering wheel cars, the U.S. manufacturers did not; this pres-
ents a serious inconvenience to driving an American car in Japan. So, why do not the
U.S. auto manufacturers make the right-hand side steering wheel cars? Is it technologi-
cally difficult for them? No, not at all. It is merely the U.S. auto manufacturers’ attitude
(not the customer-oriented, or arrogant to insist) that the customer should use their
design!
b. The Japanese toilet facility does not always have a shower set, thus a special personal
hygiene system such as this is convenient. When the restroom possesses bath shower and
toilet facilities together like in American homes, “Washlet” may not be necessary.
From the above examples, it becomes clear that, in order to expand business into a foreign country,
we need to learn the culture of that country or find a partner in that country.
By the way, TOTO “Washlet” is a sophisticated system using a smart material, a shape memory
alloy. The nozzle part of the water jet mechanism is composed of a heater and a nozzle angle control
mechanism made of a shape memory alloy. Only when the water is within a suitable temperature
range, the shape memory alloy will redirect the nozzle to the appropriate angle. When the water is
cold, the water jet is angled downwards into the flush pot.
b. Military or civilian?
Product development is sometimes supported by the government for military applications.
The researcher must understand the various differences between military and civilian
commercialization philosophies. For military applications, production quantities are rela-
tively small (several hundreds to thousands of pieces), manual fabrication processes are
generally utilized, leading to high prices. The strategy of targeting military customers may
be adopted by a small venture company as it is starting up.
The difference between required specifications and quality control is also very interesting.
Figure 11.17 shows the basic trends in quality control for military use and mass-consumer
products. Due to manual production, the production quality distribution is wider for mili-
tary products. However, all the products need to be checked for military products. No
check is required for mass-consumer products, keeping prices low. To do this, the standard
deviation of the production quality must be very small. The reader should notice that too
high a quality of the products is also “NG (not good).”
Let us consider Toshiba light bulbs. Toshiba is one of the largest light bulb suppliers in
Japan, where some of my former graduate students are working. Light bulbs typically have
an average lifetime of around 2000 h. Their quality control curve has a standard devia-
tion of ±10% (1800 ∼ 2200). If some of the production lots happen to be of a little better
quality lifetime of 2400 h, what will happen? A company executive might say bank-
ruptcy of the division. For this kind of mature industrial field, total sales amount is almost
saturated, and this 10% longer lifetime translates directly to a 10% decrease in annual

Production Production
NG OK Quality NG OK NG Quality
“All check for military use” “No check for mass-production”
—United States —Japan
FIGURE 11.17 Difference between basic trends in quality control for military use and mass-consumer
products.
income. Therefore, “too high quality” must be eliminated for mass-consumer products.
The researcher needs to understand that just seeking high quality is not the only goal of
manufacturing company. Of course, Toshiba has the technological capability to extend
bulbs’ lifetime. Toshiba does sell 2400 h lifetime bulbs. However, the price is exactly
10% higher than the usual 2000 h bulbs.
A final comment: sometimes, even famous Japanese consumer-product companies may
contribute to military/governmental applications such as the NASA Space Shuttle pro-
gram. The main reason is to obtain an aura of high quality for the company’s products,
leading to very effective advertisements, although the development effort will not bring
significant profit directly.
To tell the truth, the situation is not actually simple. Toyota, Honda reached six sigma
quality control in their automobile production, while GM and Ford are still struggling
around 3–4 sigma quality production. The key reason exists in the loyalty of the line work-
ers to the company; that is, the bottom-up employment relationship between the manage-
ment and the workers in Japan.
c. Catch the general social trends
The market also exhibits trends reflecting cultural characteristics; hence it may gradually
or drastically change with time. We consider here changes in the Japanese market trends,
which must be fully understood before an industry can expand its market globally. A sum-
mary is shown in Table 11.5. Japanese people use four Chinese character words to express
these trends, as inserted in Table 11.5.18
When the author was a university student in the 1960s, the most popular departments
at my university were metallurgy (for manufacturing steel plates and ships) and electri-
cal engineering (for building power plants), aimed at producing bigger products (heavier,
thicker, longer, and larger). However, in the 1980s, most Japanese industries became
primarily involved in electronics and computer hardware seeking device miniaturization.
My specialty, piezoelectric actuators, positioners, and USMs have been utilized to realize
the highest degree of fabrication accuracy. So, the 1980s keywords were lighter, thinner,
shorter, and smaller, all of which are completely opposite to the 1960s keywords.
In the 2000s, the keywords for new products seem to be beautiful, amusing, tasteful,
and creative, according to Dr. Hirashima. A good example is the Nintendo Game Boy,
which was initially a video game system for TVs. It has become popular worldwide among
kids (even among Japanese adults below 50 years old). Nintendo used to be a company that
supplied Japanese traditional playing cards. At the beginning of the 1970s, when most of
the Japanese electronic industries were chasing the U.S. technologies in semiconductor
devices, a major semiconductor company had a large number of imperfect 8-bit chips (the
Japanese technology at that time had such low quality). Since most of the basic functions
of those chips were active, Nintendo decided to purchase them at a very low price, and
used them (only living part of the chip) to develop computer-aided toys. The prototype

TABLE 11.5
Japanese Market Trends Over Time
1960s Heavier−Ship manufacturing
Thicker−Steel industry
Longer−Building construction
Larger−Power plant (dam)
1980s Lighter−Printer, Camera

Thinner−TV, Computer
Shorter−Printing time, Communication period
Smaller−“Walkman,” Air conditioner
2000s Beautiful−Well-known brand apparel

Amusing−TV game
Tasteful−Cellular phone (private communication)
Creative−“Culture” center, Made-to-order shoe
Game Boy did not utilize any advanced technologies, but utilized cheap 8-bit chips with
well-known technologies. The key to this big hit was its ability to fit a social trend, “amuse-
ment,” and to firmly attract the kids’ attention. Since these kids are now adults around 50
years old now, they are introducing these toys to their kids. Nintendo is now working for
two generation amusement, using the leading edge information technologies. Nintendo is
a “trend catcher,” rather than a “high-technology” company.
Table 11.6 illustrates the above-mentioned trend with the 1998 “Best Hit (Top Ten)
Products” in Japan. Except for the “flat screen TV” which is a technology-oriented (thin-
ner) product, most of others are related with “beautiful,” “amusing,” “tasteful,” and “cre-
ative” goods. In 1998, SONY was not a computer company. It was technologically a second
rank manufacturer. However, these big hits are mainly due to additional factors such as
sophisticated design. Similarly, the iMac, is a totally different product concept than the
IBM PC. SONY’s “Play Station” and Apple’s iPod, iPhone are the extensions of their suc-
cessful products. We can emphasize again that the business success does not rise merely
from the technology, but from the “trend catching” capability. These companies already
caught the new century’s consumer trend in the end of the last century.
11.3.4.2 Narrow Your Focus

After choosing a suitable customer, start narrowing development focus. The following summarizes
a procedure for narrowing the focus.
a. List all the possible application fields

When we invented piezoelectric actuators, we considered initially various application
fields:
1. Office equipment (Printer, Fax machine)

2. Cameras
3. Automobiles
Can you identify the development “pecking order” among these application areas? The
author will introduce so-called strategic or managerial decision making procedure in the
simplest way, using this practical example.

TABLE 11.6
Best Hit Products in 1998 in Japan
1. VAIO Note 505 (SONY) (Thinner, beautiful)
Silver-metallic personal
2. iMac (Apple) (Tasteful)
Inexpensive, sophisticated designed computer
3. Pocket Board (NTT Docomo) (Amusing)
Mobile telephone designed for ladies
4. Pocket “Pikachu” (Amusing)
“Tamagochi” (computer pet) and walking distance
counter
5. New Compact Car (Smaller)
Further compact size due to the Japanese automobile
regulation change
6. Draft Beer (Kirin) (Amusing)
New taste
7. Fine Pix 700 (Fuji) (Creative)
Better resolution
8. Foreign Banks
Better
9. Flat Tube TV (Thinner)
Less depth
10. Viagra (Amusing)
?
b. Start with the simplest specifications

Among the possible applications, we tried to find the simplest technological specifications.
Basically, we considered restrictions by the Japanese Industrial Standards (JIS), first.
Temperature range—The standard temperature requirements for office equipment is
between −20°C ∼ 120°C. For cameras, even though they are used outdoors, they are typi-
cally held in the hands. Thus, the temperature is always maintained around 0°C ∼ 40°C.
Much above this temperature, the film will be damaged before the camera’s failure. On
the other hand, requirements for automobile applications cover a much broader range:
−50°C ∼ 150°C.
Durability—The standard requirement for the lifetime of office equipment such as
printers is continuous operation for more than 3 months or 1011 cycles. For cameras, it is
only 5 × 104 cycles. Imagine how many pictures you take in a year. A 36-exposure roll of
film may take months to use. Automobile applications usually require durability of more
than 10 years.
Other specs may include various tests such as mechanical shock/drop, humidity, and air
pressure (applications in air plane and in space).
In conclusion, the sequence for starting development will be
Camera > Office Equipment > Automobile
As we expected, piezoactuators were first widely commercialized in a camera automatic
focusing mechanism by Canon and in a shutter by Minolta. Then, they were employed in
dot-matrix (NEC) and ink-jet (Seiko EPSON) printers. Since piezo-MLAs have been used
in diesel injection valves (automobile) by Siemens in 2000, we can say “the piezoelectric
actuator development is in a maturing period.”
It is notable that specs for cameras dramatically changed after shifting from film to
digital. Digital cameras do not use film or mechanical shutters, leading to much more

TABLE 11.7
Scoring Table for Devices
Device A Device B
Costs
1. Raw materials cost 0 1 2 0 1 2
2. Fabrication cost 0 1 2 0 1 2
3. Labor cost (special skill) 0 1 2 0 1 2
Performance
4. FOM 0 1 2 0 1 2
5. Lifetime 0 1 2 0 1 2
Market
6. Design 0 1 2 0 1 2
7. Production quantity 0 1 2 0 1 2
8. Maintenance service 0 1 2 0 1 2
Total score (add the scores) _____ _____
severe specs. The standard durability is not 5 × 104 cycles anymore, because users started to
take 10 times more pictures. The cyclic lifetime specs automatically increased by a factor
of 10 with digital cameras.
c. Consider the cost performance

We occasionally use a scoring sheet to identify a development target. A sample of how
to score is shown in Table 11.7. This table includes various factors which are significant,
including financial factors (market and cost) and device performance. We compare the
total scores and select the higher priority for development (“pecking order”).

Consider the dot-matrix printer which was developed in 1982 by NEC. Adopting a scoring
table like the one shown in Table 11.7, compare the cost performances of a bimorph and a ML
structure.
Hint
Since the deformation from the piezoelectric ceramic is small (only 1 μm), we need to amplify
the displacement for practical applications. The starting piezoceramic has a thin plate design
because a reasonable electric voltage can generate a large electric field. By bonding a piezoceramic
plate onto a metallic plate, small strain is converted to a large bending deformation (bimorph). On
the other hand, by laminating multiple thin plates, we can increase the displacement in pro-
portion to the number of the laminates (ML), which is the author’s popular patent. Refer to
Figure 11.18.21
Solution
Selection criteria for scoring are summarized in Table 11.8: (a) ML structures need large amounts
of expensive electrode materials, and the tape casting requires equipment investment. On the
other hand, fabrication is almost automatic. (b) Quick speed, high force and longer lifetime, which
are essential to dot-matrix printers, are possible with ML devices. (c) The fabrication process
of MLAs (tape casting) is most suitable for mass production. We adopted the ML piezoelectric
actuator (score 9 in comparison with 7 for Bimorph) for this printer development (Figure 11.19).

Bimorph
Multilayer
FIGURE 11.18 Two typical piezoactuator designs: Bimorph and ML.
TABLE 11.8
Example of the Scoring Table for Bimorph and MLAs
(Printer Application)
Bimorph ML
Cheap costs
1. Raw materials cost 0 √1 2 √0 1 2
2. Fabrication cost 0 √1 2 √0 1 2
3. Labor cost (special skill) 0 √1 2 0 1 √2
High performance
4. FOM 0 √1 2 0 √1 2
5. Lifetime √0 1 2 0 1 √2
Good market
6. Design 0 √1 2 0 √1 2
7. Production quantity 0 √1 2 0 1 √2
8. Maintenance service 0 √1 2 0 √1 2
Total score 7 9
FIGURE 11.19 Dot-matrix printer head using 24 ML piezoelectric actuators (1986, NEC). (From Uchino, K.
and Giniewicz, J.R., Micromechatronics, Dekker/CRC, New York, 2003.)

If the application is for cheaper devices such as ink-jet printers or cellular phones,
cost should be emphasized. Thus, the weighted score method will be applied. Suppose
that cost:performance:market weights are 4:1:1, the total scores for bimorph and ML are:
4 × 3 + 1 + 3:4 × 2 + 3 + 4 = 16:15. Hence, the bimorph actuator is recommended.
11.3.4.3 Dominate Your Market

After identifying the target, develop the products according to the following technology and product
planning creativity considerations. At the same time, consider a suitable advertising plan and price
range.
a. Advertisement (promotion strategy)

Naming or selecting a suitable trademark for a device is very important. When the author
developed cofired MLAs, they were initially named “displacement transducers.” Of course,
this is not a bad name from a physics point of view. However, it was not attractive to cus-
tomers. The name “positioned” was also used in the mechanics fields. After discussing
this with colleagues at NEC Corporation, the terminology “piezoelectric actuator” was
selected, half of which is familiar to electrical engineers (“piezoelectric”), and the remain-
ing half of which is familiar to mechanical engineers (“actuator”). Only people working in
this interdisciplinary field can understand the full meaning of this name, making it highly
suitable for a device that will be used in an interdisciplinary field.
b. Determine the appropriate price

The profit ratio for a particular sales price depends on the industry category: electronics
industries have relatively high profitability such as 10% in electronic components and 30%
in videotapes, as compared with 3%–4% for chemical commodities. On the basis of these
profit margins, we can estimate the maximum raw materials’ cost, labor costs, etc. Refer to
the rough price calculation presented in Table 11.9.
When a company is thinking about starting MLA production, they need to consider whether a
tape-casting system really needs to be installed or not. The author usually recommends the installa-
tion of a tape-casting system if the production amount exceeds 1 million pieces per year. Otherwise,
the conventional cut-and-bond method should be used by hiring several manufacturing technicians.
Also, when a company considers purchasing a new robot for automation production, consider the
price. A typical one-task robot costs $30,000, which can be used for 2 years without high main-
tenance fee. On the other hand, an annual salary of $3000 is enough to hire one worker in some
countries such as Thailand and Turkey. A manufacturing line having 10 workers corresponds to a
robot. So, an alternative solution to purchasing a robot is to start a factory in one of these countries.
The keys to be learned for this business expansion include break-even analysis, fixed cost, variable
cost, and linear programming knowledge.
TABLE 11.9
Price Calculation Sample in the Piezoactuator Area
Commercial/retail price 100
(must be comparable to competitive things)
Manufacturer’s price 50
(varies depending on the distribution channel)
Direct materials (raw materials) 10
Direct labor (engineers) 10
Manufacturing Overhead 20
Gross profit 10

Total cost
Slope = raw material cost
T
Equipment
cost Slope = raw material cost
+ labor fee
Production quantity
FIGURE 11.20 Total cost calculation comparison for the ML product among a tape-cast equipment auto-
matic production and a cut-and-paste manual production.
Figure 11.20 illustrates the manufacturing cost calculation processes for the ML actuators with
tape-cast automatic equipment and with a cut-and-bond manual production process. For the tape-
casting equipment, the initial investment (fi xed cost of equipment = $300K) is expensive, with a
low slope of variable cost (just raw materials cost) as a function of production quantity. On the other
hand, the cut-and-bond process requires a steep slope of variable cost provided by the labor fee
in addition to raw materials cost. We can find an intersection between these two lines (point T in
Figure 11.20). This product quantity is the threshold above which the equipment installation starts
to provide a better profit.
Because the labor fee is expensive in a developed country, introduction of a tape-casting facility
is usually recommended, when the production quantity exceeds only 0.1 million pieces per year.
However, due to the lower labor cost in a developing country such as Thailand and China, this
production threshold quantity is dramatically high, about 2 million pieces in the MLA production.
If the production quantity is 1 million pieces, the manual production process in these countries with
line-worker’s labor is actually cheaper. This is the globalization motivation for a corporation to set
their factory or find an OEM partner in a foreign country.
11.3.5 BUSINESS STRATEGY

The year 2008 marked the 30th year after the piezoactuator boom started. The basic technolo-
gies required seem to be maturing and a drastic cost reduction period has started. It is time to find
the strategies to move forward to wide commercialization of the piezoactuator as the key electro-
mechanical component in the twenty-first century. In general, taking into account the difficulty
in overcoming the reliability requirements such as operating temperature range and lifetime, we
discussed the development pecking order in the previous section, like “Camera” > “Office equip-
ment” > “automobile.” Since the MLA has been already utilized for the diesel injection valve, we
can say that “the piezoelectric actuator technology development has already gotten into its maturing
period.” Further, “the industrial competition has started in sharing the market by reducing the actu-
ator price drastically,” which is not only related with the manufacturing technology development
(cheaper and more efficient manufacturing), but also with the industrial managerial and strategic
decision namely the Cash-Cow product period, according to the Boston Consulting Group product
lifecycle analysis.22

11.3.5.1 Analysis Tools

From these senses, there are various constraints for commercializing piezoelectric actuators. We
will discuss these key factors, taking into account the several recent product developments. We
will adopt standard MBA tools; SWOT (Strength–weakness–opportunity–threat) Analysis and
Porter’s Five Force Model, in this section. The details can be learned in the author’s new book,
Entrepreneurship for Engineers from Taylor & Francis.23
Figure 11.21 shows the basic concept of SWOT Grid. The horizontal axis is a measure of Strength–
Weakness of the firm (your company), while the vertical axis is a measure of Opportunity–Threat
of the industries (piezoactuator area). The former is called internal environment analysis, and the
latter is called external environment analysis. All the companies are positioned in one of these four
domains, where the reader can find the suggested strategy which those companies should take.
Though the piezoactuator industries have a better Opportunity in general (Domain I or II), if the
competitive EM motor took a larger share in a particular device, the external environment becomes
a “Threat” (Domain III or IV). STEP (Social/cultural, technological, economic, and political) four
forces are considered as external factors in this chapter.
Figure 11.22 shows the concept of Porter’s Five Force model for internal environment analysis.24
First, we will start from this internal environment analysis followed by the external environment
analysis.
Opportunities
“Revitalization”
of the industries “Offensive”
developing the desired business
technology in this in this territory
territory business
Weaknesses Strengths
of the firm of the firm
“Defensive”
“Diversification”
how to overcome
Threats various product
the business in
of the industries lines, customers
this territory
FIGURE 11.21 General concept of the SWOT grid.
Development of
substitute products
Bargaining power Rivalry among Bargaining power

of suppliers competing firms of consumers
Entry of new
competitors
FIGURE 11.22 Michael Porter’s Five Force model of competition.

11.3.5.2 Internal Environmental Analysis

a. Rivalry among competing firms
Figure 11.23a shows the head structure of the Epson Mach ink-jet printer (see Section
7.7.4 for the mechanism), here the so-called cut-and-bond technology was initially used by
the supplier, Philips, to manufacture the PZT unimorph.25 Three years later, NGK, Japan
developed the cofiring technology to manufacture the PZT unimorph on the zirconia ink
chamber substrate, and took over the supplier position from Philips. The cofiring technol-
ogy provided a dramatic reduction of the manufacturing cost of the printer head, in addition
to the performance improvement; that is, the unnecessary vibration of the ink chamber was
suppressed. As the reader may know, Philips piezodevice division was closed soon after
losing this large customer, Epson. This episode teaches us the importance of continuous
effort on developing a new technology, because every product has a lifecycle.
b. Development of substitute products

Though the responsivity of a piezoactuator in the Epson printer head (Figure. 11.23a) is
much faster than the competitive Canon bubble-jet printer, the resolution of the piezo-
unimorph printer head is lower than the bubble-jet type, because the unimorph covering
area is rather large. In order to realize much finer nozzle arrangement, Epson developed
the piezo-ML type as shown in Figure 11.23b.25 This arrangement provides finer resolu-
tion and quicker speed. Though it is costlier than the unimorph type, this development is
strategically important to compete with Canon’s printers.
Actuator part
Electrode
PZT
Electrode 0.5 mm
Vibration plate
Cavity plate
Plate through hole
Adhesive layer Ink supply hole
Stainless plate
Ink chamber
Nozzle
(a)
Outer electrode (–)

Piezo element
Outer electrode (+)
Vibrate plate
Ink pass
Ink Inlet
Ink chamber
Nozzle opening
(b)
FIGURE 11.23 Piezoelectric ink-jet printer heads: (a) conventional unimorph-type and (b) new ML-type.
(From Kurashima, N., Proc. Machine Tech. Inst. Seminar, MITI, Tsukuba, Japan, 1999. With permission.)

FIGURE 11.24 Comparison between the piezo-USM (1.5 mm) and the EM motor, with a similar power
level.
The most competitive substitute of the piezoelectric actuator is the EM motors. Figure
11.24 compares the Pennsylvania State’s micro-USM (1.5 mm × 4 mm long) with an EM
motor (used in a cellular phone) with a similar output power level (50 mW). Note the signif-
icant difference in volume and weight by a factor of ×20. Also 10 times higher torque and
the efficiency of the USM are superior to the conventional EM motor. Though the USM is
a “Threat” to the EM motors in terms of performance at present, historically lower price
of the EM type is a big “Threat” to the piezoactuator.
c. Entry of new competitors

With expanding the applications and market of the piezoactuators, original raw-mate-
rial suppliers (up-stream industry) such as cement, chemical, and even steel companies
moved into the piezoactuator business. On the other hand, original customers (down-
stream industry) such as computer, home, and office electronics companies started to
manufacture the actuators. Prosthetic arm manufacturers and measuring equipment com-
panies already started to make USMs, which may be “threats” to the present piezoactua-
tor industry. However, this might be a good opportunity to find collaborative business
partners.
d. Bargaining power of suppliers

The internal electrode material of choice of the conventional ceramic ML devices was Ag/
Pd, as it was in capacitors. However, owing to the Russian economy crisis in the late 1990s,
palladium price increased dramatically with a peak in 2000, 10 times higher than 1990
(Figure 11.25). Because of this raw material cost change, capacitor industries started to use
Ni (base metal) internal electrode. Though the price of Pd is stabilized now, the effort on
the cheaper material usage is continuing. This is an example how the “Threat” should be
transformed to an “Opportunity.” The necessity creates the technology. The piezoactuator
industry is a little behind to shift to use widely the Cu internal electrode. The piezoin-
dustries are still using Ag/Pd or Pt for the internal electrodes. EPCOS recently started to
commercialize the Cu-embedded ML actuators for the diesel injection valve application,
in which the actuator cost is the major bottleneck at present.26 Refer to Section 4.2.2 for the
technical review.

Palladium—London PM Fix 1992—present

1200
1000
800
$ per ounce
600
400
200
www.kitco.com
0
Jan 92
Dec 92
Dec 93
Dec 94
Dec 95
Dec 96
Dec 97
Dec 98
Dec 99
Dec 00
Dec 01
Dec 02
Dec 03
Dec 04
Dec 05
Dec 06
Dec 07
Dec 08
FIGURE 11.25 Palladium price change with year.
e. Bargaining power of consumers

The consumer’s power depends on the application area. In the IT/office equipment applica-
tion area, the customers of piezoactuators are giant electronic companies, who have strong
bargaining power against the component companies; as such, we are a price taker. Thus
the current development target is how to reduce the production cost to meet the customer’s
desire. ML actuators and USMs need to be manufactured for much less than $3 per piece
in these days for cellular phone applications. For the robotic applications, monitoring how
the nanotechnology is expanding, as such, how nanopositioning is demanded will be the
key to estimate the future market size. Consecutively the market size will determine the
actuator price (performance/cost-oriented). In the bio/medical area, the price is not a very
important factor, but the specs (merely performance) such as very low drive voltage and
confined size are critical in designing. The piezoactuator community is fortunately a price
fi xer in this area; in other words, we can take “prestige pricing” policy. Finally, envi-
ronmental business is obviously expanding these days. Accordingly, the demand for the
actuators/transducers in the area is quickly increasing. Since it is still in the prototype
production stage, there is no threat from customers (“skimming pricing” period), but the
actuator manufacturer needs to obtain the attraction from the customers.
11.3.5.3 External Environmental Analysis

a. Economic forces
The bailout program even for private companies such as GM, Ford, and Chrysler, has just
started in the United States. Though this is against for the normal Capitalism theory, the
President took this Socialistic policy for supporting the companies not to create significant
layoffs due to the bankruptcy. The economical situation makes a company to seriously
consider how to create profit under this constraint.
We considered in the previous section the manufacturing cost calculation processes for
the ML actuators with tape-cast automatic equipment and with a cut-and-bond manual
production process. For the tape-casting equipment, the initial investment is expensive,
with a low slope of variable cost (just raw materials cost) as a function of production quan-
tity. On the other hand, the cut-and-bond process requires a steep slope of variable cost
provided by the labor fee in addition to raw materials cost. Thus, we concluded that if the
production quantity is 1 million pieces, the manual production process in these countries

with line-worker’s labor is actually cheaper. This is the globalization motivation for a cor-
poration to set their factory or find an OEM partner in a foreign country.
However, the United States is now facing another serious dilemma; that is, the weakest
U.S. dollar currency exchange rate, originated from the excess expenses with the war and
the bailout to financial companies. The import cost of the products manufactured in the
global outsourcing factories has increased, and the necessary profit cannot be created easily.
Small business corporations may need to concentrate on the Federal Research Funds tem-
porarily, restricting the R&D and manufacture of domestic products during the economi-
cal recession.
b. Political/legal forces
The twenty-first century is called “The Century of Environmental Management.” We are
facing serious global problems such as the accumulation of toxic wastes, the “Greenhouse
Effect of Earth,” contamination of rivers and seas, lack of energy sources, oil, natural gas,
etc. In 2006, the European Community started RoHS (restrictions on the use of certain
hazardous substances), which explicitly limits the usage of lead (Pb) in electronic equip-
ments. Basically, we may need to regulate the usage of PZT, the most famous piezoelectric
ceramic, in the future. Japanese and European community may experience governmental
regulation on the PZT usage in these 10 years. Pb (lead)-free piezoceramics have started to
be developed after 1999. We discussed on bismuth compounds (bismuth-layered type and
(Bi,Na)TiO3 type) and perovskite (K,Na,Li) (Nb,Ta,Sb)O3 ceramics in Section 3.2.4. RoHS
seems to be a significant threat to piezoelectric companies, who have only PZT piezocer-
amics. However, this is an opportunity for the company which is preparing alternative
piezoceramics to exchange the piezoelectric device share.
Diesel engines are recommended because by replacing the regular engine cars the puri-
fication energy can be saved, which was contributing to global warming problem. However,
the conventional diesel engine generates toxic exhaust gases such as SOx and NOx. In order
to solve this problem, new diesel injection valves were developed by Siemens with piezo-
electric ML actuators. Figure 11.26 shows a common rail-type diesel injection valve.27
c. Technological forces
Some researchers are working to reduce the drive voltage by reducing the layer thickness
of piezoelectric MLAs. However, it is not recommended for the large ML actuators for
diesel injection valve control applications in diesel automobiles. As we discussed in Figure
11.15 in Section 11.3.3, the ML actuator price increases on reducing its drive voltage (i.e.,
reducing its each layer thickness), while the driver cost increases with increasing the volt-
age. The minimum total cost is thus obtained around 160 V, leading the layer thickness of
Piezoelectric actuator
Control valve
Displacement
amplification unit
Injector body
Nozzle
FIGURE 11.26 Common rail-type diesel injection valve with a piezoelectric MLA. (From Fujii, A., Proc.

80 μm. This “sort-of standard” thickness can be derived from the cost minimization prin-
ciple, but not from the technology. Note that 10 μm layer thickness is not difficult techno-
logically in ML actuators nowadays.
We are approaching the Optical Communication era. However, compared with electric
components, optical components such as fiber couplers, optical switches are expensive.
Targeting 1/10 lower price, The Penn State Univ is developing simplest, inexpensive optical
couplers, by adopting linear USMs.28 The idea is to adopt the cofiring technique for whole
optical switch, including a base package (low temperature cofiring ceramic, LTCC) and a
linear motor (low temperature cofire PZT). Two optical fibers are inserted into this pack-
age, and the fiber position is aligned by using a Δ-shape linear motor (see Figure 11.27).
Note again that “the necessity creates the technology”; the technology which can reduce
the manufacturing cost drastically is an important research topic in these days.
d. Social/cultural forces
Though piezoelectric transformers were used on trial in the color TVs in 1970s, serious
problems were found in the mechanical strength (collapse happened in the device) and in
heat generation in color TVs, leading to the termination of production for two decades.
However, recent laptop computers with a liquid crystal display requiring a very thin, and
EM-noise free transformer to start the glow of a fluorescent back-lamp have accelerated the
development after 1990s. This is a good example how the social demand creates the technol-
ogy development (needs-push).
The best technology device does not mean the bestselling device. Bestselling is related
with the consumer attitude and social trends. Table 11.5 tabulates the Japanese consumer
attitude changes with time. Compared to the technological trends in 1980s, lighter, thinner,
shorter and smaller, we can say that beautiful, amusing, tasteful and creative will be the
keywords in 2000s.18 The present bestselling cellular phones are based on their sophisti-
cated function and artistic design, rather than the technology. Artistic sense should be added
in the device development (Sci-Art 29) in the twenty-first Engineering Renaissance era.
Competitive strategy
It is important to understand Mr. Akio Morita’s comment; “The U.S. people are focusing only on
technological creativity. But there are two more creativities; product planning creativity and mar-
keting creativity, which are equally important for commercial success.”
Last two comments from the author:
Market Research—From the market research result for 80 Japanese component industries in
1992, tiny motors in the range of 5–8 mm are highly required in these 10 years for office and factory
automation equipment, and the conventional EM motors are rather difficult to produce with suffi-
cient energy efficiency. USMs whose efficiency is insensitive to size are superior in this micromotor
area. Regarding the linear actuators, the customers’ demands are simple: 100 μm stroke, 100 N force
with 100 μs response.
Michael Porter’s Competitive Strategy—Porter’s competitive strategies24 include (1) cost leader,
(2) differentiation, and (3) focus, categorizing successful companies into four domains illustrated
in Table 11.10. NEC-Tokin is still one of the leading companies in the piezoactuator area. They are
targeting small ML actuators for cellular phone applications (focus) with mass-production facility
at the lowest cost (cost-leader). EPCOS and Denso are targeting large ML actuators for the diesel
injection valve application. These are “cost leaders.” On the contrary, Physik Instrumente is taking
a “differentiation” strategy with customer-oriented prestige price setting. Their product lines are
rather wide based on a reasonably large company size. Cedrat and Noliac are taking also differentia-
tion strategy, but keeping relatively concise product lines, with focused applications and customers.
The competitive strategies are also different depending on the application areas. “Cost-leader”
is definitively required in the information, office equipment, and the robotics area. On the con-
trary, the strategy of “differentiated product” with prestigious price is the norm in the bio- and

FIGURE 11.27 (a) 8 mm long Δ-shape linear motor stator fabricated by the LTCC compatible method at
900°C. (b) Assembled Integrated Fiber Alignment Package (IFAP™). (Courtesy by PSU; From Park, S.-H.
et al., Jpn. J. Appl. Phys., 47, 313, 2008.)
medical engineering field. The development of a focused target product is now essential in eco-
logical and energy applications. It is now your company’s choice on which development strategy
should be taken.
Chapter Essentials
1. Applications of ferroelectrics:
1. High-permittivity dielectrics
2. Ferroelectric memories

TABLE 11.10
Porter’s Competitive Strategies, and the Example Corporations,
Corresponding to the Categories
General Target Focused Target
Cost leader Murata NEC-Tokin EPCOS, Denso Nihon
Ceratec
Differentiation Physik Instrumente Cedrat
Noliac
3. Pyroelectric devices
5. Electro-optic devices
6. PTC materials
7. Composite materials
2. Present market shares of ferroelectric devices—$2 billion
1. Capacitors
3. Thermistors
3. Reliability issues of ferroelectric devices:
a. Reliability of ceramics: Reproducibility of ceramics, temperature characteristics, elec-
tric field and stress dependence of properties, aging effect
b. Reliability of devices: Electrode materials, electrode designs, layer thickness depen-
dence, failure detection techniques
c. Drive techniques: Pulse drive method, heat generation mechanism, high-power drive
technique
4. Bestselling devices:
a. Business strategy: Technology, product planning, and marketing creativities (after
A. Morita)
b. Technological creativity: Serendipity, analogy, product effect
c. Product planning creativity: Seeds and needs, developing speed, specifications
d. Marketing creativity: Choose your customers, narrow your focus, and dominate your
market
e. Analysis tools—Strength–Weakness–Opportunities–Threats (SWOT) analysis:
Internal environment analysis—Michael Porter’s Five Force model
External environment analysis—STEP (social/cultural, technological, economic, and
political/legal forces)
5. Directions of smart systems:
a. Adding components for higher function
b. Reducing components for miniaturization and lower cost
Check Point
1. (T/F) PTC thermistor may be one of the most promising research areas of ferroelectric
devices at present. True or False?
2. (T/F) Developing an efficient manufacturing process requires less creativity than discover-
ing a new material. True or False?
3. (T/F) Marketing does not require any creativity of the engineer, and should be transferred
merely to the sales expert. True or False?

4. (T/F) The drive voltage of piezoelectric MLAs should be reduced as low as 12 V for auto-
mobile applications, because the car battery is 12 V. True or False?
5. (T/F) Once one device development fails, a similar development may not restart in 25–30
years, which is known as the R&D cycle. True or False?
6. A prosthetic arm company started to manufacture piezoelectric USMs in their factory.
What is this situation categorized among the Michael Porter’s Five Force model?
7. What are four words abbreviated as “STEP” in the external environment analysis of the
industry?
8. (T/F) When Lexmark develops slightly modified piezoelectric inkjet printers against
Epson’s piezoelectric printers, this is called “development of substitute products.” True or
False?
9. (T/F) When Canon develops bubble-jet printers against Epson’s piezoelectric ink-jet print-
ers, this is called “development of substitute products”. True or False?
10. Why may the most famous PZT be legally restricted in the future?
Chapter Problems
11.1 Your company is requested to manufacture 1000 pieces of prestigious price ML actua-
tors ($1000 each) per year. How will you manage this request? (1) manufacturing
manually by hiring manpower in the United States, (2) purchasing a manufacturing
equipment (tape-cast robot) in the United States, or (3) manufacturing manually by hir-
ing manpower in Thailand through a partner company, or else?
11.2 Your company ordered 1000 pieces of ML actuators to a Japanese manufacturer for
Japanese Yen 10,000 each piece on October 1st (currency exchange rate: $1 = JY100).
You received the products on November 1st (currency exchange rate: $1 = JY90), and
need to pay immediately this invoice. How much did your company obtain or lose
the money during this 1 month in the U.S. dollar amount, regarding this international
trade?
11.3 Discuss the superiority of your product, such as cost, performance, and market,
in comparison with the competitive products? Use a “scoring table” and discuss it
quantitatively.
EXAMPLE
Suppose your company developed a micro-USM called “metal tube type,” consisting of a metal
hollow cylinder and two PZT rectangular plates actuators. Raw materials cost is lower than
the conventional EM motor’s thin copper wire. The metal cylinder motor 2.4 mm in diameter
and 12 mm in length was driven at 62.1 kHz in both rotation directions. A no-load speed of
.
1800 rpm and an output torque up to 1.8 mN m were obtained for rotation in both directions
under an applied rms voltage of 80 V. Higher maximum efficiency of about 28% and torque
for this small size compared with the EM type is a noteworthy feature. The key features of
this micromotors are (1) world-smallest motors with human finger fighting torque level (much
superior to the EM motors) and (2) low manufacturing cost competitive with the lowest EM
motors. Table 11.11 compares the superiority scores for the metal-tube and EM motors for cel-
lular phone camera applications. Miniature size, high torque, and low cost are essential factors
for the choice. The metal-tube type has advantages in FOM and small size, in comparison with
the EM micromotors.
11.4 “Don’t read papers?”—My PhD advisor was Professor Shoiichiro Nomura at Tokyo
Institute of Technology, Japan. He taught me first “Don’t read papers,” when I joined his
laboratory. I also use it when I teach my graduate students. What is the real meaning of
this?

TABLE 11.11
Scoring Table for Comparing Metal-Tube USM and EM Motors
(Cellular Phone Application)
Metal Tube EM
Costs
1. Raw materials cost 0 1 √2 0 √1 2
2. Fabrication cost 0 √1 2 0 1 √2
3. Labor cost (special skill) 0 √1 2 0 √1 2
Performance
4. FOM 0 1 √2 √0 1 2
5. Lifetime 0 1 √2 0 1 √2
Market
6. Design (small size) 0 1 √2 √0 1 2
7. Production quantity 0 √1 2 0 1 √2
8. Maintenance service 0 √1 2 0 √1 2
Total score 12 9
Comment
I had top academic grades during my undergraduate period. I read many textbooks and academic
journals. Accordingly, whenever Professor Nomura suggested that I study a new research topic,
I said things such as “that research was done already by Dr. XYZ, and the result was not promis-
ing …” After having a dozen of these sort of negative conversations, partially angrily, partially dis-
appointedly, Professor Nomura ordered “Hey, Kenji! You are not allowed to read academic papers
for a half year. You should concentrate on the following experiment without having any biased
knowledge. Having a strong bias, you cannot discover new things. After finishing the experiment
and summarizing your results, you are allowed to approach the published papers in order to find
whether your result is reasonable, or is explainable by some theory.” Initially, I was really fearful
of getting totally wrong results. However, I finished it. That led to my first discovery: PMN–PT
electrostrictive materials.
Remember that knowing too much suppresses innovative work. A real discovery is usually made
by a young less-experienced engineer. Once he/she becomes an expert professor, unfortunately
he/she loses some creativity.
REFERENCES
1. J. Ceram. Soc. Jpn., December issue (1984).
2. J. Ceram. Soc. Jpn., December issue (1990).
3. K. Uchino: Piezoelectric Actuators and Ultrasonic Motors, Kluwer Academic Publishers, Boston, MA
(1996).
4. K. Uchino: Proc. 9th Int’l. Symp. Appl. Ferroelectr., p. 319 (1995).
5. K. Abe, K. Uchino, and S. Nomura: Jpn. J. Appl. Phys., 21, L408 (1982).
6. K. Nagata: Proc. 49th Solid State Actuator Study Committee, JTTAS (1995).
7. K. Uchino and H. Aburatani: Proc. 2nd Int’l Conf. Intelligent Materials, p. 1248 (1994).
8. H. Aburatani and K. Uchino: Am. Ceram. Soc. Annu. Mtg. Proc., SXIX-37-96, Indianapolis, IN, April
(1996).
9. J. Zheng, S. Takahashi, S. Yoshikawa, K. Uchino, and J. W. C. de Vries: J. Am. Ceram. Soc., 79, 3193
(1996).

10. N. Kanbe, M. Aoyagi, S. Hirose, and Y. Tomikawa: J. Acoust. Soc. Jpn. (E), 14(4), 235 (1993).
11. A. Morita: Private communication (1986).
12. http://en.wikipedia.org/wiki/Benjamin_Franklin#Early_life (as of March, 2008).
14. R&D Innovator, 4(3), Winston J. Brill & Associates, Redmond, WA (1995).
15. K. Uchino, M. Yoshizaki, K. Kasai, H. Yamamura, N. Sakai, and H. Asakura: Jpn. J. Appl. Phys., 26(7),
1046–1049 (1987).
16. Battelle Company Report (1995).
17. A. M. Flyn, L. S. Tavrow, S. F. Bart, R. A. Brooks, D. J. Ehrlich, K. R. Udayakumar, and L. E. Cross:
J. Microelectromech. Syst., 1, 44 (1992).
18. Y. Hirashima: Product planning in the feeling consumer era, Jitsumu-Kyoiku Publ., Tokyo (1996).
19. C. R. Thomas and S. C. Maurice: Managerial Economics, McGraw-Hill Irwin, New York (2005).
20. M. Treacy and F. Wiersema: Discipline of Market Leaders, Addison-Wesley Publishing, Reading, MA
(1996).
21. K. Uchino and J. R. Giniewicz: Micromechatronics, Dekker/CRC, New York (2003).
22. D. A. Aaker: Developing Business Strategies, John Wiley & Sons, New York (1984).
23. K. Uchino: Entrepreneurship for Engineers, Taylor & Francis, Boca Raton, FL [in press].
24. M. Porter: Competitive Strategy: Techniques for Analyzing Industries and Competitors, Free Press, New
York (1980).
25. N. Kurashima: Proc. Machine Tech. Inst. Seminar, MITI, Tsukuba, Japan (1999).
26. F. Boecking and B. Sugg: Proc. New Actuator 2006 (Bremen, June14–16), A5.0, p. 171 (2006).
27. A. Fujii: Proc. JTTAS Meeting Dec. 2, Tokyo (2005).
28. S.-H. Park, J. Agraz, S. Tuncdemir, Y.-D. Kim, R. E. Eitel, A. Baker, C. A. Randall, and K. Uchino: Jpn.
J. Appl. Phys., 47(1), 313–318 (2008).
29. M. Fusakul: Proc. 5th Int’l Conf. Intelligent Mater. (ICIM2003), State College, PA, June 14–17 (2003).

Index
A magnetoelectric effect, 311
need-pull force, 304
Accelerometer, 175 piezoelectric ceramic/polymer composites,
Acoustic impedance, 166 308–309
Active fiber composite (AFC), 289 Schottky barrier, 314
Actuator materials, 198 seed-push market, 304
Alkoxide method, 86 smart materials, 308
Alkoxides, hydrolysis of, 87 Bimorph
Antiferroelectricity structure, 98, 176, 208
antiferroelectrics, 51–52 transient vibration of, 202–203
dielectric and elastic equations, 53 Binary systems, 168
Kittel’s free energy expression, 52 Biosensor, 193–194
Antiferroelectric state, 238 Blanchard grind, 285
Antiresonance state, 182–183 BL capacitor, see Barrier-layer capacitor
Bulk electro-optic devices
B eye protection application, 243
ferroelectric picture memory device, 242–243
Banno’s modified cubes model, 282–283 KTN single crystal optical switch, 251–252
Barium titanate (BaTiO3) laser beam scanner, 252
crystal structures, 17 stereo TV application, 243–244
Curie–Weiss law, 17–18 two-dimensional displays
dielectric and piezoelectric properties of, 168–169 fabrication process of, 244–246
direct precipitation method, 86 image projector, construction of, 247–249
grain, 263 light valve array, characteristics of, 246–247
high-permittivity capacitor, 115 optical display, 244
ionic displacements, 7 BURned-out Plastic Spheres method (BURPS), 283
lanthanum, 261 Burn out binder, 285
modified with dopants, 169 Business strategy
particle size, function of, 105 analysis tools, 328
phase transitions, 19 external environmental analysis
piezoelectric, temperature dependence, 104 economic forces, 331–332
semiconductive ferroelectric ceramics, 75 political/legal forces, 332
strain curve, electric-field, 103 social/cultural forces, 333
temperature dependence of, 106 technological forces, 332–333
Barrier-layer capacitor, 274 internal environmental analysis
Bearingless rotor flexbeam, 207 Canon bubble-jet printer, 329
Bestselling devices competitor entry, 330
marketing creativity consumer bargaining power, 331
choosing customers, 320–322 ink-jet printer, 329
market domination, 326–327 supplier bargaining power, 330
narrowing focus, 322–326
product planning creativity
C
antiaging products, 315
development pace, 316 Camera auto zooming/focusing mechanism,
2D PLZT display, 314 218, 221
electromechanical vibrations, 315 Capacitor, types, 116
innovation obstacle, technology management, 316 Cathode ray tubes, 246
managerial economics, 318 Cellular phones
piezoelectric transformers, 314 1-D atomic model, 22
smart systems, 318–319 Gordon Moore’s Law, 23
specifications, 316–317 Ceramic actuators
technological creativity barium stannate, 199
electromagnetic (EM) effect, 311 designs for, 200–201
figure of merit (FOM), 304 electric field-induced strains in, 199
flying kite experiment, 305 modified PZT, 198–199
functionality matrix concept, 310 pulse-drive techniques for, 202
339

340 Index
Ceramic capacitors grain size, 104

capacitance of, 92 vs. temperature, 128
classes of, 115 Diesel fuel injection timing chart, 211
Ceramic powders Differential-type piezoelectric printer head
ferroelectric, 87 element, 209
piezoelectric, 99 Diphasic composites, 280
preparation of Direct piezoelectric effect, 161
alkoxide hydrolysis, 86–87 Disk-type hula-hoop motor, 223–224
coprecipitation, 86 Displacement amplification mechanism, hinge
solid-state reaction, 85–86 lever, 98
Ceramics Domain contributions
BaTiO3, 103 crystal structure-coercive field, 61
Dy-doped, 103 domain engineering
fabrication processes perovskite ferroelectric, 65
ceramic powders preparation, 85–87 PZT films, 66
single crystal growth, 90 hysteresis estimation program
sintering process, 87–90 Boltzmann probability, 63
templated grain growth, 90–91 Gibbs energy curves, 62
plate-shaped piezoelectric, 99 P–E hysteresis curve, 63
CFRP/PZT cantilever beams, 295 reorientation, 54–57
Chip capacitors electric field-induced strain, 54
base-metal internal electrode polarization, 56
barium titanate ceramics, 117 polycrystalline state, 55
PbTiO3-based chip, 117–119 Uchida−Ikeda model
thin multilayer capacitor, 116–117 PLZT ceramics, 61
Chopped IR irradiation, pyroelectric response, 147 polarization, 58
Cofiring, 92 reorientation, 57–58
Cole–Cole relation, 127 Domain wall pinning model, 80
Color TV system projection, 247 Doping, PZT system (PbZrxTi1−xO3), 170
Combination effect, 275 Dot-matrix printer head, 203, 208–209
Composition fluctuation, computer simulation of, 122 Double-minimum potential, Cole–Cole plot for, 127
Converse piezoelectric effect, 8, 161, 176 Double-minimum potential ion, 125
Core-shell model, 107 2-D PLZT optical display
Cross talk test system, 249 driving circuit for, 248
CRT, see Cathode ray tubes fabrication process, 245
Crystal deficiency model, 81 matrix segment, 244
Crystal structure and ferroelectricity, 1–3 3-D positioning ML actuator, 200–201
Crystal structure models, 121, 130 DRAM, see Dynamic random access memory
Curie temperature, 75–76, 128, 239 3-D stress sensor, 174
Curie–Weiss law, 17–18, 123, 264–265 D-TGS crystal, IR image target of, 154
Curved linear array, 188 Dynamic equation
Cut-and-bond method, 92 for electrical circuit, 185
Cylindrical gyroscope, 175 for mechanical system, 184
Cymbal energy harvesting composites, 226 Dynamic random access memory, 131
capacitor films, 136
ferroelectric, 135–136
D principle of, 131–134
structure of, 132
Damped capacitance, 183
DC–DC buck converter, 226
Device designing E
flexible composites, 99
flextension/hinge-lever amplification mechanisms, Efficiency, 164
98–99 Electric field
multilayer structures, 92–94 and external stress, 161
single disk devices, 91–92 temperature changes, 158
thin/thick films vs. induced strain curve, 163
constraints, 102 Electric field-induced strain
film manufacturing techniques, 99–101 electrostriction
MEMS applications, 101–102 microscopic explanation, 8
unimorph/bimorph devices, 94–98 PMN-based ceramic curve, 9
Dichroic mirrors, 247 piezoelectric strain
Dielectric constant microscopic explanation, 8
capacitor dielectrics, 75 PZT-based ceramic curve, 9

Index 341
Electric polarization EM coupling factors, 179

charge accumulation and frequency dependence, 3 Energy harvesting devices
microscopic origins, 2 backgrounds of, 284
Electrocaloric devices, electrocaloric effect piezoelectric composite, 288–290
electrocaloric materials, 157–158 Energy transmission coefficient, 162–165
phenomenological formula, 154–157 EO effect, see Electro-optic effect
Electrocaloric effect, 156
electrocaloric materials, 157–158 F
phenomenological formula, 154–157
Electrocaloric materials, 157 FEM, see Finite element method
Electromagnetic (EM) motors, 152, 212 FeRAM, see Ferroelectric RAM
vs. metal-tube ultrasonic motor, 337 Ferpic, see Ferroelectric picture memory device
vs. piezo-ultrasonic motor, 330 Ferroelectric ceramic devices
Electromechanical coupling, 156, 284 electro-optic properties of, 77
Electromechanical coupling factor, 162–164, fabrication of, 75
177–178 Ferroelectric ceramics
Electromechanical parameters calculation, 183–184 multilayer structures, 92
Electronically controlled shock absorber, 210 nonlinear polarizability of, 235
Electro-optic effect, 11–16 Ferroelectric film
birefringence, 235 polarization vs. electric field curve, 136–137
brightness, 247 thin polyvinyldifluoride (PVDF), 152
development of, 239 Ferroelectricity
ferroelectric ceramic application of
(Pb,La)(Zr,Ti)O3 system, 236 ceramic capacitors, 115–116
Pb(Zn1/3Nb2/3)O3, 241–242 chip capacitors, 116–119
PMN–PT, 239–241 hybrid substrates, 119–121
polycrystalline microstructure of, 236 relaxor ferroelectrics, 121–125
half-wave voltage, 235 size effect on
K(Ta,Nb)O3, 251 3-D particle, 105–108
light shutter, 258 grain size effect, 103–105
construction of, 236 nanotechnology, 102
second-order, 235 Ferroelectric picture memory device, 242–243
optical parametric devices, 235 Ferroelectric powders synthesis, 87
PLZT composition, relation between, 236 Ferroelectric/pyroelectric phase, 151
primary/secondary EO coefficient, 237 Ferroelectric RAM, 137
scanner development of, 140
hybrid scanner device, 252 flash vs. DRAM, 141
scanning frequency of, 252 Ferroelectric relaxors, multipotential well
voltages, seven spots of, 253 model, 124
waveguides, 253–254 Field effect transistors, electron–hole pair
Electro-optic materials, 75 generation, 131
Electrostriction Field-induced strain, dopant effect, 81
converse effects, 50 Finite element method simulation software, 290
electric field-induced strain Flextensional piezoelectric actuator, 98
microscopic explanation, 8 Flight actuator mechanism, dot-matrix printer head
PMN-based ceramic curve, 9 using, 203
phenomenology, 47–50
temperature dependence, 50–51 G
Electrostrictive actuators, 201–202
Electrostrictive bimorph-type stress sensor, 176 Gas igniter, 174
Electrostrictive coefficients, matrix notation Gate voltage
axis electrostrictive strain, 41 drain current, 139
crystallographic point groups, 35 electron inversion layer, 132
electrostrictive coefficient matrix, 38–40 I D vs. EDS curves, 134
nonlinear-linear algebra conversion, 35 Gaussian distribution, 241
piezoelectric coefficient matrix Gaussian error distribution, 122
centrosymmetric point group, 36 Gauss’s law
noncentrosymmetric point group, 36–37 domain wall stability, 80
piezoelectricity, 34 head-to-head polarization configuration, 78
piezoelectric shear strain, 41 GBL capacitor, see Grain boundary layer
Poisson’s ratio, 40 capacitor
shear deformation, 40 Gibbs energy, 76
Electrostrictive effect, 9 Gordon Moore’s Law, 23
Electrostrictor, 80 Grain boundary, 261

342 Index
Ba1−xLaxTiO3, 263 Lead lanthanum zirconate titanate (PLZT) material,

n-type semiconductive grains, 264 313–314
Grain boundary layer capacitor, 269–270 Lead magnesium niobate
core-shell model, 270 high-resolution electron-microscope image of, 123
model of, 269 PT single crystals, SSCG technique, 90
type of, 269 temperature dependence of, 124
Green sheets, preparation steps, 93 Lead zirconate titanate, 75, 85
crystal deficiencies, 81
H crystal growth of, 90
domain-wall pinning model, 81
Half-wave voltage, 243 electrodes, 137
High power energy harvesting, 225–227 electromagnetic devices, 292
High-power ultrasonics, 189 fabrication process, 100
“Hubble” telescope, 207 films, fatigue of, 138
Hula-Hoop motion, 217 heat generation monitoring, 156
Human genome mapping, 315 lead-containing ceramics, 81
Hybrid scanner geometry, layout of, 252 micromachining process, 101
Hybrid substrates, 119–121 micropump, 101–102
Hydrolic reaction, 86 multilayer actuators, 110
Hysteresis, 82, 200 oxygen deficiencies, 80
Hysteresis estimation program permittivity/electromechanical coupling factor, 76
Boltzmann probability, 63 phase diagram, 77
Gibbs energy curves, 62 piezoelectric properties, 78
P–E hysteresis curve, 63 piezoelectric response, comparison of, 278
polarization vs. electric field curves, 142
polymer composites, 99, 279
I powder, 0-3 composite material, 281
powder, relative permittivity plot, 281
Impedance curves, 184
rhombohedral, epitaxially grown, 101
Infrared image sensors, 152–154
rods, 1-3 composite of, 278
Ink-jet printer, 208
silicone composite, 279
Intelligent actuator system, 302
single-phase piezoelectric polymers, 277
Intuitive crystallographic model, 121
soft/hard properties, 79
Irradiation chopping frequency, 149
spurrs epoxy composite, 279
thin films of, 158
J vibration velocity, 83
mechanical quality factor vs. mole
Joule heating resistor, 284 fraction, 84
vs. temperature rise, 83
K Light intensity, 236
Light waveguides, 252
Kerr effect, 259 Linear array transducers, 188
Kittel’s free energy expression Linear array-type ultrasonic probe, 188
antiferroelectrics, 52–54 Liquid crystal displays, 246
electrostrictive coupling, 52 Load mass model, 163
K(Ta,Nb)O3 (KTN) Longitudinal mechanical vibration
crystal optical switch module, 251 current flow, 182
optical switch module, 251 extensional, 180
C-band, performance of, 252 stress, electric field, and induced strain, 181
single crystal, low cleavage of, 251 through transverse piezoelectric effect, 181
Lotgering’s factor, 91
L Low-energy harvesting, 227
Luminous flux, 247
Landau theory
first-order transition, 45
free energy, 43 M
polarization, 43
second-order transition, 44–45 MACH ink-jet printer head, 210
Langevin vibrator, 216 Macro fiber composite, 227, 288
Latching relay, 200 output voltage signal, 291
Layer-structure ferroelectrics, 138 smart material corporation, 289
LCD, see Liquid crystal displays stress distribution, 290
Lead-based relaxor materials, 172 Macroscopic domains, 128

Index 343
Magnetic field, 276 MFMOS structure, see Metal-ferroelectric-MOS

Magnetic irradiation, 290 structure
Magnetoelectric device, 292 Michael Porter’s competitive strategy, 333–334
Magnetoelectric effect, 276 Michael Porter’s five force model, 328
Magnetoelectric material, 276 Microchip capacitors, 115
Magnetoelectric sensors, 290–291 Microelectromechanical systems (MEMS), 315
Magnetostrictive CoFe2O4, 276 Micromass sensors
Magnetostrictor disks, 290 biosensor, 193–194
Magnetron sputtering apparatus, principle of, 100 viscosity sensor, 194
Market research, 333 Microrotary motors, 216–217
Market share Microscopic composition fluctuation model, 122
ferroelectric devices MMC substrate, see Monolithic multicomponent ceramic
actual sales amount, 298 substrate
Japanese manufacturing companies, 299 MOCVD, see Metal-oxide chemical vapor deposition
piezoelectric actuators, 298–300 Monolithic multicomponent ceramic substrate,
Mason’s equivalent circuits, length expander bars, 196 119–121
Material designing cross-sectional view, 120
composition selection, 75–77 dimensions of, 117
fabrication processes, 75 voltage-controlled oscillator, 120
ferroelectricity, dopant effects on Moonie/cymbal actuator
crystallographic deficiencies, 80–82 end cap design, 99
domain wall stability, 78–80 piezoelectric ceramic, 98
electro-optic properties, 77 structures of, 99
high-power characteristics, 82–84 Morphotropic phase boundary (MOP), 76, 169, 236
Mathematical treatment electric field, 77
electrostriction potential minimum wells, illustration of, 79
converse effects, 50 rhombohedral–tetragonal, 83
phenomenology, 47–50 MOS, see Metal-oxide-semiconductor
temperature dependence, 50–51 Motional capacitance, 183
ferroelectric domain contributions Multilayer actuators
crystal structure-coercive field, 61 rail-type diesel injector, 212
domain engineering, 64–67 walking piezomotor, 211
hysteresis estimation program, 61–64 Multilayer ceramic actuators
reorientation, 54–57 domain texture creation model, 112
Uchida–Ikeda model, 57–61 electroded portion of, 94
tensors fabrication process, 93
crystal symmetry and tensor form, 27–33 structure of, 92
electrostrictive coefficients, matrix notation, 34–42 two domain texture models, 112
fundamentals, 25–26 Multilayer ceramic speaker, 186
reduction, 33–34 Multilayer piezoelectric element, 200–201
representation, 26–27 Multilayer thermistor PTCR effect, 268
Mechanically free sample, admittance for, 182 Multilayer-type transformer, 195
Mechanical quality factor, 166 Multiple element array transducers, 188
Mechanical resonance frequency, 202
Mechanical sector transducers, 188 N
Mechanical system
dynamic equation, 184 Nonvolatile ferroelectric memory
and electrical, equivalency between, 185 characteristic degradation, categories of, 137–138
Metal-ferroelectric-MOS structure, 139 FERAM, 136–138
Metal-ferroelectric-semiconductor FET (MFSFET) MFSFET, 139–140
drain current vs. gate voltage relation, 139 NTC–PTC effect, 276
structure of, 139
Metallic sheet, 94 O
Metal-oxide chemical vapor deposition, 100
Metal-oxide-semiconductor, 131 (Pb,La)(Zr,Ti)O3
energy band model, 131–133 DC poling, 243
SiO2 film, 135 dielectric relaxation, 104
Metal-oxide-semiconductor FET (MOSFET) electric field polarization and birefringence, 237
drain/source voltage, n-channel changes, 134 EO coefficients, grain size dependence of, 238
p-type semiconductor, 133 EO shutter, 237
SiO2 film capacitor, 131 glasses, stereo TV system, 244
structure of, 132 grain growth, 89
MFC, see Macro fiber composite grain size dependence, 104–105

344 Index
microphotographs of, 88 Piezocomposites, 173

optical display, 244 advantages of, 273
piezoelectric application, 75 composite effects
precipitation of, 86 combination effect, 275
stereo TV, eye glasses, 243 product effect, 275
transmittance of, 245 sum effect, 275
Oxide-mixing technique, 85 types of, 274
two-phase composites, 274
Piezoelectric actuators
P application fields, 322
applications
Parallel bimorph, 209 pulse-drive motors, 208–212
Pb-free piezoceramics, 20 servo displacement transducers, 207–208
Pb(Zn1/3Nb2/3)O3 ceramic actuator materials, 198–200
birefringence vs. electric field, 241 classification, 198, 201–202
coefficients for, 77 designs of, 200–201
domain reversal mechanism, 242 flextensional, 98
lattice parameters, 78 market share, 298–300
microscopic domain reversal mechanism, 242 TEMS with, 211
normal/abnormal grain growth, 88 Piezoelectric BaTiO3, 276
phase diagram, 77 Piezoelectric bimorph, 95
temperature dependence, 78 Piezoelectric buzzer, 186
Pb(Mg1/3Nb2/3)O3-PbTiO3 system Piezoelectric ceramic cylinder, 217
EO coefficient, changes, 240 Piezoelectric ceramic plates, unimorph/bimorph
refractive index, 240 devices, 94
transmittance of, 239 Piezoelectric ceramic powders, 99
PbTiO3 Piezoelectric ceramics, 174
chloroprene rubber 0-3 composites, 282 Piezoelectric coefficient matrix
coefficients for, 77 centrosymmetric point group, 36
lattice parameters, 78 electrostrictive coefficients
Ni/Cu, thermodynamic calculation, 120 centrosymmetric point group, 36
phase diagram, 77 noncentrosymmetric point group, 36–37
temperature dependence, 78 noncentrosymmetric point group, 36–37
Pen-ferroelectic region, 236 Piezoelectric composite materials
Percolation threshold, 287 composite effects
PFC, see Piezo fiber composite combination effects, 275–276
Phase-change-related strains, 199 product effects, 276–277
Phase modulation, 253 sum effects, 274–275
Phase transition, sharp/diffuse/successive, 160 0-3 composite material, 281–283
Phenomenological theories 3-3 composites, 283–284
antiferroelectricity connectivity, 273–274
antiferroelectrics, 51–52 piezoceramic:polymer composite, 277–278
dielectric and elastic equations, 53 principle of, 278–281
Kittel’s free energy expression, 52 useful for, 273
electrostriction, 47–50 Piezoelectric constant, 281
ferroelectricity Piezoelectric cylinder torsion vibrator, 218
fundamentals, 42–43 Piezoelectric dampers, 284
Landau theory, 43–45 backgrounds of, 284
Photoconductive film, 242–243 composite dampers, 284–288
Photodetector, 249 electrical energy, transformation, 285
Piezoactuators, 208 patch, 287
Piezoceramic composite, 277 Piezoelectric devices
Piezoceramic ferroelectric hysteresis, 280 longitudinal length extension modes, 178
Piezoceramic fiber, 278 planar extensional vibration modes, 178
Piezoceramic patch, 224 Piezoelectric effect
Piezoceramic plate, longitudinal mechanical converse, 161
vibration, 204–205 direct, 161
current flow, 182 Piezoelectric/electrostrictive ceramics, 86
dynamic equation, 202–203 Piezoelectric energy harvesting system
extensional, 180 high-energy harvesting, 225–227
overshoot height, 206 low-energy harvesting, 227
stress, electric field, and induced strain, 181 piezoelectric passive damping to, 224–225
through transverse piezoelectric effect, 181 with switch shunt, 226
Piezoceramics, 187 Piezoelectric equations, 177

Index 345
Piezoelectric fibers, 289 Polyurethane rubber, 281

Piezoelectricity Polyvinyldifluoride, 172
direct, 176–177 pyroelectric IR sensor, 152
meaning, 161 structure, 173
Piezoelectric lead zirconate titanate (PZT), 18–19 Polyvinylidene difluoride (PVDF), 19, 277
Piezoelectric materials, 167, 284 Positive temperature coefficient
piezocomposites, 173 demonstration of, 266
piezoelectric properties of, 168 dopants, 261
polycrystalline materials experimental verification, 265–266
barium titanate, 168–169 honeycomb air heater, hair dryer, 266
PZT system, 169–172 schottky barrier effect, 263–265
PVDF, 172–173 semiconducting properties, 261–263
relaxor ferroelectrics, 172 thermistor exhibits, 267
single crystals, 168 thermistors, temperature-change detection,
thin films, 173–174 266–268
Piezoelectric resonance, 182 Positive temperature coefficient of resistivity (PTCR)
and antiresonance states, 183 effect, 75
electromechanical coupling factor, 177–178 BaTiO3
mode, equivalent electrical circuit, 84 ceramics, 262
piezoelectric equations, 177 grain boundary, 263
Piezoelectric resonators and EM coupling factors, 179 thermistors, 262
Piezoelectrics ceramics
application, 75 grain size, 268
figure of merit impulse current change, 271
acoustic impedance, 166 over-current/voltage protector, 271
electromechanical coupling factor, 162–164 Pressure sensors, 174
mechanical quality factor, 166 Product planning creativity
piezoelectric strain constant, 161 antiaging products, 315
piezoelectric voltage constant, 161 development pace, 316
power density measurement, figures of merit, 82 2D PLZT display, 314
vibration velocity, 82 electromechanical vibrations, 315
Piezoelectric scanner, 252 innovation obstacle, technology management, 316
Piezoelectric single crystals, 192 managerial economics, 318
Piezoelectric strain constant, 161–162 piezoelectric transformers, 314
Piezoelectric strips, bearingless rotor flexbeam, 207 smart systems, 318–319
Piezoelectric transformer specifications, 316–317
definition, 194 Projection-type color TV system, 250
fundamental structure, 194–195 Propagating-wave type motors, 215–216, 223–224
ML-type transformer, 195 linear motor
Piezoelectric vibrators, equivalent circuits of, 185–186 using bending vibration, 220
Piezoelectric voltage constant, 161–162 vibrators, 219
Piezoelectric wafers, 284, 287 rotary motors
Piezo fiber composite, 289 PZT piezoelectric ring, 221
Piezo Toyota electronic modulated suspension, 210 vibration source, 220–221
PLZT, see Lead lanthanum zirconate titanate; (Pb,La) vs. electromagnetic motor, 221–222
(Zr,Ti)O3 PT, see PbTiO3
PLZT light valve PTC, see Positive temperature coefficient
electrode configuration, 246 Pulse-drive techniques, 202
projection-type color TV system, 250 PVDF, see Polyvinyldifluoride; Polyvinylidene difluoride
schematic electrode configuration of, 246 Pyroelectric coefficient, 157
PLZT:PVDF composites, 288 Pyroelectric detectors
PLZT thin film optical waveguide geometric configurations, 146
electrode configurations, 257 room-temperature properties, 150
optical switch, 255 Pyroelectric devices, 145
SEM picture of, 254 Pyroelectric effect
PMN electrostrictor, 207 principle on, 145
PMN-PT system, see Pb(Mg1/3Nb2/3)O3-PbTiO3 system pyroelectric component, 148
Poisson’s ratio, 40 Pyroelectric IR detector, amplifier for, 149
Polycrystalline ceramics, fabrication processes, 85–91 Pyroelectric materials
Polycrystalline ferroelectric ceramics, crystallographic composite structures for, 294
texturing of, 90 figure of merit, 150–152
Polycrystalline materials pyroelectric effect, 145–148
barium titanate, 168–169 responsivity, 148–149
PZT system, 169–172 Pyroelectric sensors, 150, 153

346 Index
infrared (IR) radiation, 145 motor characteristics of, 222

principle of, 146 stator structure of, 222
temperature monitoring, 145 Schottky barrier, 263
transducing optical/thermal energy to electrical Seiko’s motor, 223
energy, 150 Semiconductor substrate
Pyrovidicon tube diffusion, 136
disadvantages of, 153 Nb:SrTiO3, 254
images, dark night, 155 optical waveguide fabrication, 254
structure of, 152–153 Sensor electrode, 95
PZN–PT single crystal, 188 Serendipity, 304–305
PZT, see Piezoelectric lead zirconate titanate Servo displacement transducers, 207–208
PZT ceramic ring, 221 Shape memory alloys
PZT system (PbZrxTi1−xO3) phase transition analogy, 309
doping, 170 water jet mechanism, nozzle, 320
energy conversion rate in, 172 Shim, see Metallic sheet
lead titanate, 170 Single crystal
phase diagram, 169 growth
piezoelectric, dielectric, and elastic properties PZN–PT, PMN–PT, and PZT, 90
of, 171 quartz, LN, and LT, 90
PZT thin films, 174 materials, 168
preparations, 75
Q Single-crystal relaxor ferroelectrics, 172
Sintering, 87–88
Quartz, 168 Skanavi-type dielectric relaxation, multipotential well
model, 124
R Smart materials, 1
Smooth impact drive mechanism (SIDM), 224
Rail-type diesel injector with piezoelectric ML actuator, Sol-gel method, 86
212 Solid-state displacement transducers, 198
Rattling ion model, see Intuitive crystallographic model Solid-state single crystal growth, 90
Rayleigh wave, 191 Sonochemistry, 188–189
Reactive-templated grain growth, 91 Spaghetti syndrome, 318–319
Relaxor ferroelectrics, 172 Spontaneous polarization, 4–6
dielectric relaxation, 124–125 SSCG, see Solid-state single crystal growth
diffuse phase transition, 122–123 S-shape mode bimorph, 208
high permittivity, 121–122 Standing-wave type motors, 223
permittivity of, 123 linear motor
Reliability issues components, 218–219
aging effect, 301 walking principle, 219
drive/control techniques, 303 rotary-type motor
electrodes, 301 microrotary motors, 216–217
field-induced polarization, 300 torsion coupler, 216
intelligent actuator system, 302 two-vibration-mode coupled type
lead magnesium niobate (PMN) material, 301 motor, 218
safety systems, 303 Strength–weakness–opportunity–threat (SWOT)
strain gauge configuration, 302 analysis, 328
zero-point drift, 301 Strontium titanate
Replamine method, 283 BST, 136
Resistive shunt method, 224 dielectric constant, film thickness, 135
Resistive shunt piezoelectric patch, 225 Surface acoustic wave (SAW)
Resistivity, 261 actuator applications, 193
Rhombohedral symmetry phases, 76–77, 79 filter, 191
Ring-type bimorph, 208–209 generation, 192
Room temperature distillation, 190 material parameters for, 192–193
Rosen-type transformer, step-up voltage ratio for, technology, 191
195–197 transducers, 191
RTGG, see Reactive-templated grain growth velocity, 192
Rubber composite coaxial cables, 283 Surface acoustic wave devices
fundamental structure of, 191
S materials used for, 191
Surface wave coupling factor, 192
Sashida’s motor Suspension system, 211
for camera automatic focusing mechanism, 221 Switch shunt, 226

Index 347
T Trifluoroethylene (TrFE), 173

Triglycine sulfate (TGS), 145
Tape-casting method, 92 Two-vibration-mode coupled-type motor, 218
Technological creativity
electromagnetic (EM) effect, 311
figure of merit (FOM), 304
U
flying kite experiment, 305 Uchida–Ikeda model
functionality matrix concept, 310 PLZT ceramics, 61
magnetoelectric effect, 311 polarization, 58
need-pull force, 304 reorientation, 57–58
piezoelectric ceramic/polymer composites, Ultrasonic imaging
308–309 with PZT ceramic probes, 189
Schottky barrier, 314 transducer for, 187–188
seed-push market, 304 Ultrasonic motor (USM)
smart materials, 308 applications, 82
Temperature distribution, rectangular plate, 83 basic construction of, 212–213
Temperature/infrared sensors, 152 classification of
Templated grain growth, 90 propagating-wave type, 215–216
TEMS, see Piezo Toyota electronic modulated standing wave type, 214–215
suspension comparison of, 223–224
TEMS-adaptive suspension system, 211 demerits of, 214
Tensors merits of, 213–214
crystal symmetry and tensor form metal-tube USM vs. EM motors, 337
electric conductivity, 27 piezo-USM vs. EM motor, 330
piezoelectric tensor, 29 resonance mode operation, 303
symmetric tensor, 29 rotor, 213
electrostrictive coefficients, matrix notation smart system development, 319
axis electrostrictive strain, 41 torsion coupler, 216–217
crystallographic point groups, 35 Ultrasonic scanning detectors, 187
electrostrictive coefficient matrix, 38–40 Ultrasonic transducers
nonlinear-linear algebra conversion, 35 geometry, 187
piezoelectric coefficient matrix, 36–37 for ultrasonic imaging, 187–188
piezoelectricity, 34 Unimorph beam structure, 286
piezoelectric shear strain, 41 Unimorph bending actuation, 223
Poisson’s ratio, 40
shear deformation, 40
fundamentals V
coordinate system, 25 VCR head tracking actuators, 209
transformation matrix, 26 Vibration damping test, resistive shunt piezoelectric
reduction, 33–34 patch for, 225
representation, 26–27 Vibratory coupler-type motor, 214–215
TGG, see Templated grain growth Video cameras, 243
Thermal IR camera, 156 Viscosity sensor, 194
Thermodynamic phenomenological theory, 154
Thickness-mode electromechanical W
coupling, 173
Thin-film capacitors, 120 Walking piezomotor
Thin films, 173–174 with ML actuators, 211
Thin film-type electrocaloric demo kit, 158 using inchworm mechanism, 212
Thin piezoelectric ring, 221 Waveguide modulators
Thin polyvinyldifluoride ferroelectric film, 152 LiNbO3 single crystals, 252–254
Thin-/thick-film techniques, 75 PZT/PLZT thin film, 254–255
Torsion coupler, 216 Wet chemical preparation methods, 75
Transient vibration of bimorph, 203
Transparent electro-optic ceramics
ferroelectrics, nonlinear polarizability of, 235
X
(Pb,La)(Zr,Ti)O3 system, 236–238 Xenon lamp, 247
Pb(Zn1/3Nb2/3)O3, 241–242
PMN–PT, 239–241
Trapped-energy filter, 189–191 Z
Trapped-energy principle, 190
Traveling wave, 220 Zero shifts, 134
Traveling wave-type motors, 222 Zinc oxide (ZnO), 173

Ferroelectric Devices 2nd Edition

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Ferroelectric Devices 2nd Edition

Uploaded by

Document Information

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

Ferroelectric Devices 2nd Edition

Uploaded by

Copyright:

Available Formats

Ferroelectric

© 2010 by Taylor & Francis Group, LLC

Boca Raton London New York

CRC Press is an imprint of the

© 2010 by Taylor & Francis Group, LLC

No claim to original U.S. Government works

International Standard Book Number-13: 978-1-4398-0376-9 (eBook - PDF)

© 2010 by Taylor & Francis Group, LLC

Chapter 1 General View of Ferroelectrics ....................................................................................1

Chapter 2 Mathematical Treatment of Ferroelectrics .................................................................25

© 2010 by Taylor & Francis Group, LLC

Chapter 3 Designing with Materials and Devices and Fabrication Processes............................ 75

Chapter 4 High-Permittivity Dielectrics ................................................................................... 115

Chapter 5 Ferroelectric Memory Devices................................................................................. 131

© 2010 by Taylor & Francis Group, LLC

5.2 Nonvolatile Ferroelectric Memory .................................................................. 136

Chapter 6 Pyroelectric Devices ................................................................................................ 145

Chapter 7 Piezoelectric Devices ............................................................................................... 161

© 2010 by Taylor & Francis Group, LLC

Chapter 8 Electro-Optic Devices .............................................................................................. 235

Chapter 9 PTC Materials .......................................................................................................... 261

© 2010 by Taylor & Francis Group, LLC

9.1.2 Mechanism of the PTC Phenomenon................................................... 261

Chapter 10 Composite Materials ................................................................................................ 273

Chapter 11 Future of Ferroelectric Devices................................................................................ 297

© 2010 by Taylor & Francis Group, LLC

11.3.4 Marketing Creativity ........................................................................ 319

© 2010 by Taylor & Francis Group, LLC

© 2010 by Taylor & Francis Group, LLC

• FAQs (frequently asked questions)—to answer common in-class questions,

© 2010 by Taylor & Francis Group, LLC

0. Course Explanation and Prerequisite Knowledge Check 1 Time

Lab Demonstration 2 Times

© 2010 by Taylor & Francis Group, LLC

© 2010 by Taylor & Francis Group, LLC

Q3 v = 1 / ρS E [0.5 point for v = 1/ρsE]

Q10 P = dX (refer to x = dE)

© 2010 by Taylor & Francis Group, LLC

© 2010 by Taylor & Francis Group, LLC

© 2010 by Taylor & Francis Group, LLC

1.1 CRYSTAL STRUCTURE AND FERROELECTRICITY

© 2010 by Taylor & Francis Group, LLC

Input Material Output

Output Charge Magnet-

Permittivity Elect.-mag. Converse Elec. caloric Elec.-optic

Mag.-elect. Magneto- Mag. caloric Mag. optic

Piezoelectric Piezomag. Elastic Photoelastic

Pyroelectric Thermal Specific

Diagonal coupling Sensor

FIGURE 1.1 Microscopic origins of the electric polarization.

© 2010 by Taylor & Francis Group, LLC

Microwave Infrared Ultraviolet

Total polarizability (real part)

108 1012 1016

FIGURE 1.2 Frequency dependence of the polarizability (or permittivity).

σf : Free charge σb: Bound charge σt: True charge

FIGURE 1.3 Charge accumulation in a dielectric capacitor.

© 2010 by Taylor & Francis Group, LLC