______________________________________________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) electrooptic devices,
(6) PTC materials, and
(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 most cases. In the case of the light sensor, for example,
semiconductive materials are superior to ferroelectrics in response speed and
sensitivity. Magnetic devices are popular for memory applications, and liquid
crystals are typically used for optical displays. High permittivity dielectric thin film
can survive in DRAMs, but commercialization of ferroelectric memory (FRAM) is
uncertain because of the variability of the coercive field of the material.
Ferroelectric devices may fail to be developed when competitive materials already
exist. Therefore, we see again, ferroelectrics are strong only 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),

(2) Thin film hybrid sensors (pyro-, pressure, acceleration sensors), and
(3) Electrooptic devices (light wave-guides, thin film hybrid displays).

Of course, this is not meant to discount the other areas of potential development.
However, it is anticipated that these 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.

275
276 Chapter 11

11.1 MARKET SHARE

(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 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. Since
growth in these areas has been suppressed since 1992 due to the worldwide
economical recession, no noticeable change is expected until the beginning of the
next century.

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.
1.14 1.18
1.2 (1983) (1991)

others
(17%)
1.0 cylinder
(3%)
semicondu
0.8 ctor (8%)
0.35 0.61
(1983) (1991)

0.6 multilayer
(35%) US device (3%)
high-voltage 0.1 0.3
0.4 device (12%) (1983) (1991)

plate acoustic device (21%)

0.2
(37%) SAW device (28%)
c eramic filter
PTC
0 (36%)

Capacitor Piezoelectric Thermistor

Device

Fig. 11.1 Actual sales of ferroelectric devices (with respect to device type) for
1983 and 1991 fiscal years in Japan.
Future of Ferroelectric Devices 277

(b) Piezoelectric Device $0.35 B

(a) Capacitor $1.14 B

Others
(12%) Others
KCK (21%) Murata
(7%) Murata (30%)
Matsushita- (30%) TDK(3%)
Elec.Compo. Toko (5%) Matsushita-
(11%) Elec.Compo.
Taiyo- Sanyo (5%)
TDK (17%)
yuden NGK-NTK
(28%) Kyocera
(12%) (7%)
Toshiba (9%)
(7%)

(d) Total $1.6 B

(c) Thermistor $0.1 B
Others
(9%) Others
(17%) Murata
Matsushita-
Murata KCK (5%) (30%)
Elec.Compo.
(42%) Kyocera (5%)
(19%) TDK
(30%)
Taiyo-yuden
(8%) TDK
Matsushita- (22%)
Elec.Compo.
(13%)

Fig. 11.2 Market share of ferroelectric devices by Japanese manufacturing

companies (1983).

(2) Market Share of Ceramic Actuators

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 active
vibration control using relatively large actuators (> 30 cm). The piezo-actuators and
ultrasonic motors have been developed largely by private industries in Japan, aimed
primarily at applications involving precision positioners and compact motors with
tiny actuators of less than 1 cm. Research and development are too consumer-
oriented to be supported by the Japanese government. The only big national project
related to this area concerns "micromechanisms," which mainly involves the silicon
micro -machining-related micromotors. European developments are a little behind
278 Chapter 11

Japan and the United States, and they seem to have been searching for a variety of
possible applications. The device sizes at the trial manufacturing stage range are
generally around 10 cm.

The markets in the United States are limited to military and defense applications,
and it is difficult to estimate the sales amount. Among the current needs of the
Navy are: smart submarine skins, hydrophone actuators, propeller noise cancellation
devices; and of the Air Force: smart aircraft skins; while the Army requires
helicopter rotor twisting, aeroservoelastic control and cabin noise/seat vibration
cancellation devices.

In Japan, piezoelectric camera shutters (Minolta Camera) and automatic focusing

mechanisms in cameras (Canon), dot-matrix printers (NEC) and part-feeders
(Sanki) are now being commercialized and mass-produced on the order of tens of
thousands of pieces per month. Piezoelectric ink jet printers (Epson) and
piezoelectric transformers (NEC, 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.

Table 11.1 Summary of ceramic actuator developments comparing the United

States, Japan and Europe.

_________________________________________________________________

U.S. Japan Europe

_________________________________________________________________
TARGET Military -oriented Mass-consumer Lab-equipment
product product product
CATEGORY Vibration Micro-motor Micro-motor
suppressor Positioner Positioner
Vibration
suppressor
APPLICATION Space structure Office equipment Lab stage/stepper
FIELD 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 Tokin Corporation Philips

MANUFAC- Morgan Matroc NEC Siemens
TURERS Itek Opt. Systems Hitachi Metal Hoechst CeramTec
Burleigh Mitsui Chemical Ferroperm
AlliedSignal Canon Physik Instrumente
Seiko Instruments
_________________________________________________________________
Future of Ferroelectric Devices 279

The annual sales of ceramic actuator units, camera -related devices and ultrasonic
motors in 2005 in Japan are estimated to reach $500 million, $300 million and $150
million, respectively.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.

11.2 RELIABILITY ISSUES

The potential and range of application for ferroelectric materials have been
highlighted in the previous section. However, there still remain various 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 with
respect to materials, device designs, and drive/control techniques.

(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 co-precipitation and sol-gel will be required.

Porosity must be eliminated completely from the sintered ceramic, when it is used
for electrooptic devices. On the other hand, porosity does not affect the
piezoelectric strain behavior significantly, even when it is more than 94%. The tip
deflection of a unimorph made from 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 provides "soft" characteristics, the
sample exhibits larger strains and less hysteresis when driven under a high electric
field (1 kV/mm). On the other hand, acceptor doping provides "hard"
characteristics, leading to a very small hysteretic loss and a large mechanical quality
factor when driven under a small AC electric field (that is, ultrasonic motor
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 high temperature sensors and actuators for automobile engines
and cryogenic applications for lab equipment and space structures.

Systematic studies on the high electric field and stress dependences of ferroelectric
devices are also eagerly awaited, as well as the composition dependence of
mechanical strength.
280 Chapter 11

Although the aging effect is very important, not many investigations have been
done so far. The aging effect arises from two factors: depoling and 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. 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. The change in lifetime of a ferroelectric device,
such as a multilayer piezoelectric actuator, with temperature T and DC bias voltage
E obeys an empirical rule:6)

tDC = A E-n exp(W DC/ kT) (11.1)

where W DC is a sort of activation energy and n is a characteristic parameter.

(2) Reliability of Devices

The popular silver electrodes have a serious problem of migration under a high
electric field and high humidity in actuator, electrooptic and memory applications.
This problem can be overcome with the use of a silver-palladium alloy (or more
expensive Pt). To produce inexpensive ceramic actuators, we need to introduce Cu
or Ni electrodes, which requires a sintering temperature as low as 900°C. Even
though ceramics that can be sintered at low temperature have been developed for
capacitors, they have yet to be developed for actuators.

Delamination of the electrode layer is another problem for multilayer types as well
as for bimorphs. To enhance adhesion, composite electrode materials of a metal
and ceramic powder colloid, ceramic electrodes, and electrode configurations with
via- holes are recommended. To suppress the internal stress concentration which
initiates cracks in the device, several electrode configuations 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 feedback, 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 (for
example, in a lathe machine). Acoustic emission measurement of a piezo-actuator
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 piezo-
device can be utilized as an AE sensor.
Future of Ferroelectric Devices 281

Actuation
Feedback (2)
Breakdown detection
Piezoelectric sensor
actuator

Feedback (1)
Strain sensor

Control
voltage
Signal (1)

Signal (2)

Computer-controlled
power supply

Fig. 11.3 Intelligent actuator system with both position feedback and breakdown
detection feedback mechanisms.

TR6847

High voltage supply

Fig. 11.4 Strain gauge configuration of the internal electrode for an intelligent
actuator.

A special internal electrode configuration with a strain gauge configuration has been
proposed to increase the reliability of multilayer piezoelectric actuators.8) As shown
in Fig. 11.4, strain gauge configured electrode patterns are inserted at every ten
282 Chapter 11

internal layer of a multilayer actuator. 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 Fig. 11.3.

(3) Drive/Control Techniques

Ferroelectric devices generally have quick responses. However, when a sharp pulse
or step-like voltage is applied to a device, an unstable output ringing tends to occur
just after the voltage is applied. This occurs even in capacitors and electrooptic
devices, where it is sometimes called "screaming" because of the sound it
sometimes generates. It is caused by a piezoelectrically or electrostrictively induced
mechanical resonance.

In addition to the unstable output, pulse driving the ferroelectric generates very
large tensile stresses in the device, sometimes large enough to initiate cracks. 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.

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 piezoelectric transformers and ultrasonic motors.
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., surface area).9) It is necessary to select a suitable duty ratio for the drive
power so as to produce a temperature rise no greater than 40°C.

As far as high power ultrasonic transducers and motors are concerned, operation in
the antiresonance mode has been proposed.10) Ultrasonic motors have
conventionally been operated in the resonance mode, at the so-called "resonance"
frequency. However, the mechanical resonance state at the "antiresonance"
frequency reveals much higher Q and lower heat generation than observed for the
"resonance" condition. 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 resonance operation. This means that a
conventional inexpensive power supply may be utilized for driving the ultrasonic
device.

(4) Safety Systems

Future research and development should focus on ecologically superior systems

(that is, those fit for humans!) as well as technologically advanced ones. Lead
zirconate titanate based ceramics were the key players highlighted in this textbook.
However, the development of lead-free, non-toxic ferroelectric ceramics will
Future of Ferroelectric Devices 283

certainly be even more important by 2005, when we can expect some political
regulation on materials containing lead. Pb-free single crystals, such as BaTiO3
and K(Ta,Nb)O3 , will be studied vigorously in the near future, particularly in the
fields of medical and automobile applications.

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 multilayer
piezoelectric actuators is a good example of a future safety system.

11.3 DEVELOPMENT OF BESTSELLER DEVICES

The author has been closely involved with 60 ferroelectric devices for more than 25
years. During these developments, the author has been a professor, a vice president,
an R&D center deputy director or a standing auditor at several universities and
private companies both in Japan and the United States.

In this last section, the author wishes to describe his personal philosophy on how to
develop bestseller devices, especially in the area of smart materials and structures.
This sort of "how-to" is, the author believes, much more important to younger
researchers than practical knowledge about devices.

(1) Business Strategy

Mr. Akio Morita, former president of SONY Corporation, responded to a question

from a journalist concerning the lack of creativity on the parts of Japanese
researchers by saying "Japanese researchers are good at chasing and imitating the
original idea for commercialization, but they in general lack creativity." Mr. Morita
suggested that there should be three types of creativity with respect to Research &
Development 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 resolution), for example, has indeed been transferred from Philips. However,
even though the idea came from Philips, they could not commercialize it with their
supporting technologies. Matsushita, on the other hand, succeeded in
commercializing the idea after an intensive three-year development on it. Of course,
it is up to the reader (you!) to decide which company is at a higher level with respect
to science and technology. It is, however, apparent that only Matsushita obtained a
large profit from this TV development.

Table 11.2 summarizes the three important types of creativity to be implemented

when developing R&D strategy, each of which will be described in further detail in
the following sections.
284 Chapter 11

Table 11.2 Three types of creativity in research and development.

___________________________________________________________

STRATEGY IN R & D

(1) Technological Creativity - New functions

- High performance

(2) Product Planning Creativity - Specification

(sensitivity, size, power)
- Design

(3) Marketing Creativity - Price

- Advertisement
___________________________________________________________

(2) Marketing Creativity

"Discipline of Market Leaders"11) 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, and
(c) Dominate your market.

We will consider the details of this concept according to these steps.

Choose your customers

Domestic or Foreign?

Let us start by solving the following example problem.

Example Problem 11.1______________________________________

(a) Japanese cars are popular in the United States, but U.S. cars are not popular in
Japan. Can you guess 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?
Future of Ferroelectric Devices 285

Solution

(a) Initially, the reader needs to notice the traffic system difference: 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 presents a serious inconvenience to driving an American
car in Japan.

(b) The Japanese toilet facility doesn't always have a shower set, thus a special
personal hygiene system such as this is convenient. When the restroom possesses
both bath shower and toilet facility together like an American house, "Washlet" may
not be necessary.
___________________________________________________________________

From the above examples, it becomes clear that, in order to expand our business
into a foreign country, we need to learn the culture of that country or to 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 the shape memory alloy.
Only when the water is warmed up by the heater to a temperature within a suitable
temperature range, the shape memory alloy will function to redirect the nozzle to
the appropriate angle. When the water temperature is cold, the water jet is angled
downwards into the flush pot.

Military or Civilian?

Product development is sometimes supported by the government for military

applications. The researcher must understand the various differences between the
philosophies behind military and civilian commercialization. For military
applications, since the production quantity is relatively small (several hundreds to
thousands of pieces), manual fabrication processes are generally utilized, leading to
high prices for the products. 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 11.5 shows the basic trends in quality control for military use
and mass-consumer products. Due to the manual production, the production quality
distribution is wider for military use. However, all the products will be checked for
military products. No check is required for mass-consumer products, however, to
keep the price low. So, the standard deviation of the production quality must be
very small. Also, the reader will likely realize that too high quality of the products
is also "NG (not good)."
286 Chapter 11

Production Production

NG OK Quality NG OK NG Quality
“All check for military use” “No check for mass-production”
- US - Japan

Fig. 11.5 Difference between basic trends in quality control for military use and
mass-consumer products.

Let us consider as a good example Toshiba light bulbs. Toshiba is one of the largest
light bulb suppliers in Japan. The light bulbs typically have an average lifetime of
around 2000hr. Their quality control curve has a standard deviation of ± 10% (1800
- 2200). If some of the production lots happen to be of a little better quality lifetime
of 2400hr, what will happen? A company executive might mention bankruptcy of
the division. For this kind of mature industrial field, the total number of sales is
almost saturated, and this 10% longer lifetime translates directly to a 10% decrease
in annual income. Therefore, "too high quality" must be strictly eliminated for
mass-consumer products.

Of course, Toshiba has the technological capability to extend the bulb's lifetime. If
the reader has a chance to visit Japan to look for light bulbs, 2400 hr-lifetime bulbs
can be found in shops. The reader should not be surprised, however, to find the
price exactly 10% higher than the usual 2000 h r bulbs.

A final comment: sometimes, even a famous Japanese consumer-product company

may contribute to military/governmental applications such as the NASA Space
Shuttle program. The main reason for this is to obtain a certificate of high quality
for that company's product, leading to a very effective advertisement although the
development effort will not bring significant profit directly.

Catch the General Social Trends

The market also exhibits trends, which reflect cultural characteristics, and, hence,
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 in Japan. A summary is shown in Table 11.3. Japanese people
use "four Chinese character words" to express these trends, as shown at the bottom
of Table 11.3.12)
Future of Ferroelectric Devices 287

Table 11.3 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

(Projected) Amusing -- T.V. game
Tasteful -- Cellular phone (private commun.)
Creative -- "Culture" center, Made-to-order shoe
____________________________________________________________

1960s
1980s
200 0s

When the author was a university student, the most popular departments in my
university were metallurgy (for manufacturing steel plates and ships) and electrical
engineering (for building power plants), aimed at producing the bigger products.
However, once into 1980s, most of the Japanese industries became primarily
involved with electronics or computer hardware, and sought the miniaturization of
devices. Piezoelectric actuators, positioners and ultrasonic motors have been
utilized to realize the highest degree of fabrication accuracy.

In the 2000s, the keywords for new products will be "beautiful," "amusing,"
"tasteful," and "creative." A good example is Game Boy by Nintendo, which is a
video game system for TVs, and has become popular worldwide among kids.
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 failure ranked 8-bit chips (the Japanese technology
at that time had such a level!). Since most of the basic functions of those chips were
alive, Nintendo decided to purchase them at a very low price, and used them to
develop computer-aided toys. The prototype Game Boy did not utilize any
advanced technologies, but utilized the cheap 8-bit chips with well-known
technologies. The key to this big hit was its ability to fit a social trend,
"amusement," and to firmly attract the kids' attention.
288 Chapter 11

Narrow your focus

After choosing a suitable customer, we will start to narrow our development focus.
The following summarizes a procedure for narrowing the focus.

1. List all the possible application fields.

2. Start from the simplest specifications:

Among the possible applications, find the simplest technological specifications.

Practive this process with the following Example Problem.

Example Problem 11.2________________________________________________

Choose the field of application in which we can most easily utilize ceramic
actuators with typical specifications:

1. Office equipment (Printer, Fax machine)

2. Cameras
3. Automobiles

Solution

(1) Temperature range

The standard temperature requirements for office equipment is between - 20 -
120o C. For cameras, even though they are used outdoors, they are typically held in
the hands. Thus, the temperature is always maintained around 0 - 40o C. As you
can imagine, at much above this temperature, the film will be damaged before the
camera's failure. On the other hand, the requirements for automobile applications
cover a much broader range: - 50 - 150o C.

(2) Durability
The standard requirements 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 x 104 cycles. Imagine how many pictures you take in a year. A 36-exposure roll
of film installed in spring may still be used in fall! Automobile applications usually
require durability of more than 10 years.

In conclusion, the sequence for pecking order is:

(Camera> Office Equipment > Automobile).
___________________________________________________________________

3. Consider the cost performance:

We occasionally use a scoring sheet to identify the development target. A sample

of how to score is shown in Table 11.4. Compare the total scores, and select the
higher priority for development.
Future of Ferroelectric Devices 289

Table 11.4 Scoring table for devices.

Device A Device B _
Cheap Cost
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) figure of merit 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
______________________________________________________________

Example Problem 11.3________________________________________________

Considering the dot-matrix printer, compare the cost performances of a bimorph and
a multilayer structure using a scoring table like the one shown in Table 11.4.

Solution

An example scoring is as follows:

COMMENTS: (a) Multilayer 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 multilayer devices. (c) The
fabrication process of multilayer actuators (tape- casting) is most suitable for mass
production.

Bimorph Multilayer___
Cheap Cost
1) raw materials cost 0 x1 2 x0 1 2
2) fabrication cost 0 x1 2 x0 1 2
3) labor cost (special skill) x0 1 2 0 1 x2
High Performance
4) figure of merit 0 x1 2 0 1 x2
5) lifetime x0 1 2 0 1 x2
Good Market
6) design 0 x1 2 0 x1 2
7) production quantity 0 x1 2 0 1 x2
8) maintenance service 0 x1 2 0 x1 2
_______________________________________________________________
TOTAL SCORE 6 10
___________________________________________________________________
290 Chapter 11

Dominate your market

After identifying the target, we develop the products according to the following
technology and product planning creativity considerations. At the same time, we
need to consider a suitable advertisement and price range.

Advertisement

Naming or selecting a suitable trademark for a developed device is very important.

When the author developed co-fired multilayer actuators, 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 the customers. The name "positioner" 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 electricians
("piezoelectric"), and the remaining half of which is familiar to mechanics
("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.

Learn the appropriate price

The profit ratio for a particular sales price depends on the industry catagory:
electronics industries have relatively high profitability such as 10% in electronic
components and 30% in videotapes, as compared with 3 - 4% for chemical
industries. 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 11.5.

When the reader's company is thinking about starting a multilayer actuator business,
they will need to consider if a tape-casting system really needs to be installed. The
author usually recommends the installation of a tape-casting system if the
production amount exceeds 1 million pieces per year. Otherwise, the conventional
cut-and-bond method should be employed by hiring several manufacturing
assistants.

Also, when the reader's company considers purchasing a new robot for automation
production, consider the price. A typical one-task robot costs $ 30,000 (US$),
which can be used for two years without high maintenance fee. On the other hand,
an annual salary of $ 3,000 (US$) is enough to hire one worker in some counries
such as Thailand and Turkey. A manufacturing line having ten workers
corresponds to a robot. So, an alternative solution to purchasing a robot is to start a
factory in one of these countries.
Future of Ferroelectric Devices 291

Table 11.5 Price calculation sample.

______________________________________
Commercial price 100
(must be comparable to equivalent things)

Manufacturer's price 25 - 50
(varies depending on the circulation route)
Raw materials 10
& Labor cost 10
& Profit 5
______________________________________

(3) Technological Creativity

There are two different approaches in exercising technology creativity: to find a

new functional effect or material and to achieve a high performance or figure of
merit. These are typically called "research" and "development," respectively.

New function

Serendipity is often an important factor in discovering a new function in a material.

A good example can be found in piezoelectric polymer, PVDF, which Dr. Kawai
discovered accidentally. Another example the reader may know of is a high
temperature superconducting ceramic discovered by Drs. Bednortz and Muller.

We are told that every researcher has "3 lucky chances" in his/her life to discover
new things (a traditional Japanese proverb). However, most people do not even
recognize these chances and lose their chances. Only the people who are ready to
accept these can really find the new phenomenon. A Japanese company executive
mentioned that a person who develops a 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 be promoted to president. From this illustration the reader can
understand how difficult it is to develop an actual bestseller product.

The personality and aptitude of the researcher are, of course, also important factors
here. Why don't you try the following Example Problem 11.4 to assess your ability
to experience serendipity?
292 Chapter 11

Example Problem 11.4________________________________________________

First, familiarize yourself with the contents of this page as much as possible in one
minute.

Fig. 11.6 Test picture. (Note that this article was randomly cited from an
academic journal.)
Future of Ferroelectric Devices 293

Second, answer True or False for the following sentences:

(1) His name is Ohuchi.

(2) His article is printed on p.15 of an academic journal.

(3) He has moustache.

(4) He wears a dotted-design tie.

Solution

(1) F, (2) T, (3) T, (4) F.

COMMENTS:
___________________________________________________________________
Your Score Aptitude
___________________________________________________________________
Really recognized 4 You can be a good engineer.
Guess because asked so 2 - 3 You fit to a manager/sales engineer.
No idea 0 Abandon your dream to be an engineer.
___________________________________________________________________

The person who aims to be an engineer tries to remember the written content first.
If you failed to answer questions (1) and (2) correctly, you must recognize your
aptitude. The "moustache" recognition is also expected, because it belongs directly
to this person. However, some readers may not remember his tie, to which you
should give further attention. "You can see it only when you try to."

The author usually asks unconventional questions of a job interviewee to our

company, such as:

(a) You climbed up the company stairs a couple of minutes ago. How many stairs
did you climb?
(b) You must have seen a 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 to the walking direction differ remarkably. When the answer is "I don't
remember," we usually suggest he return home. Even when the answer is correct,
"left," if the answer is given as a guess and the correct answer probability is 50%, he
may be hired for a management position. Only when the correct answer arises from
a confident memory, will we hire him as a professional engineer.
___________________________________________________________________
294 Chapter 11

If you missed the above three chances, what should you do? Quit research? The
following example is dedicated to the unlucky reader, who, like the author, missed
those lucky chances. We can still research using a more systematic way, for
example, by using our intuition. The author is making use of (1) secondary effects
and (2) scientific analogy.

(1) As is well known, any phenomenon has primary and secondary effects, which
are sometimes recognized as linear and quadratic phenomena, respectively. In
electrooptic devices, the Pockels and Kerr effects correspond to the primary and
secondary effects, as you learned in this textbook. In actuator materials, these
correspond to the piezoelectric and electrostrictive effects.

When the author started actuator research in the middle of the 1970s, precise
"displacement transducers" (we used this terminology initially) were required in a
Space Shuttle program, in particular for "deformable mirrors," for controlling the
optical pathlengths over several wavelengths. Conventional piezoelectric PZT
ceramics were plagued by hysteresis 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 a centro-symmetric crystal, is not
affected by hysteresis or aging. The response should be much faster than the time
required for domain reorientation in piezoelectrics/ferroelectrics. In addition,
electric poling is not required.

However, at that time, most of the people believed that the secondary effect would
be a minor effect, 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
found that relaxor ferroelectrics, such as the lead magnesium niobate-based solid
solutions exhibit enormous electrostrictions.

(2) Probably most of the 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. 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. 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.

High performance

The concept of composite effects is very useful, particularly for systematically

improving the properties and figure of merits. As we learned in Chapter 10, a
combination effect can provide an improved figure of merit g (= d/ε) in
piezoelectric PZT:polymer composites.
Future of Ferroelectric Devices 295

Product effects are more attractive. Philips' magnetoelectric material is a good

example, which can be employed as a simple magnetic field monitor. The author's
photostrictive materials were also discovered along a similar line of reasoning. The
following anecdote cited from R&D Innovator13) will be of interest.

___________________________________________________________________
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 telephone 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 mechanical 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
micromechanisms 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
work 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 "piezoelectric" 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
zirconate 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.
296 Chapter 11

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 opposite polarization, then connected the edges. I shined a purple light to one side,
which generated a photovoltaic 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. When light alternately irradiated
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 converted 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 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 necessary; but what is necessary is listening to others outside your particular research
area: for instance, basic researchers or people with specific, applied objectives.
___________________________________________________________________

The situation of discovering "monomorphs" (semiconductive piezoelectric bending

actuators) is also similar to the above. When attending a basic conference of the
Physical Society of Japan, the author 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. A
Future of Ferroelectric Devices 297

polycrystalline piezoelectric sample was used first, and some reduction processes
were attempted to expand the Schottky barrier thickness. We succeeded finally in
developing a monolithic bending actuator. The "Rainbow" structure, recently
developed by Aura Ceramics, is one of the monomorph modifications, although the
fabrication process is their original work.

(4) Product Planning Creativity

Seeds and Needs

What the author usually suggests to a person in the product planning division in a
company is to reexamine 10-year-old research. If the social needs still exist, and
because the related patents have probably expired or will expire soon, there will
likely be a good business opportunity. Most importantly, find out the reasons for
the failure and judge your company's capability to overcome them.

Examples can be found in 2D displays and piezoelectric transformers. As described

in Chapter 8, the 2D PLZT display was realized by using the newly developed
nano-powder technology and a tape-casting method. We always need to watch
carefully for possible supporting technological development.

Piezoelectric transformers were commercialized for the first time in the beginning
of the 1970s for supplying high voltage to color TVs. However, this application
disappeared in less than a year, due to easy cracking or destruction of the device.
The success of the second commercialization owes to three key factors: strong
social demand for a back-light inverter application, matured powder technology to
provide sufficiently mechanically strong piezoceramics, and advanced design
technology such as Finite Element Method softwares to simulate electromechanical
vibrations.

Prediction of future technologies is also important in finding "seeds" for product

planning processes. Battelle's prediction (Battelle reports regularly on future
technological needs), the top ten for 2005, is listed:14)

___________________________________________________________________

1. Human genome mapping. Genetic-based personal indentification 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 molecular 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.
298 Chapter 11

4. Digital, high-definition TV. A major breakthrough for American television

manufacturers -- and a major source of revenue -- that will lead to better advanced
computer 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. Anti-aging products -- that rely on genetic information to slow the aging process --
will include aging 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
sophisticated tastes of computer-literate students.
___________________________________________________________________

Note that there is a very high possibility of using ferroelectric devices, especially in
the areas, 2, 3, 4, 5, 6, 8 and 9.

Further down-sizing of actuators will be required in medical diagnostic applications

such as blood test kits and surgical catheters. Silicon Micro Electro-Mechanical
Systems (MEMS) have currently been developing rapidly. However, electrostatic
force itself is, in general, too weak to move something with sufficient mechanical
efficiency. Piezoelectric thin films compatible with silicon technology will be much
more focused upon micro-electromechanical systems. An ultrasonic rotary motor as
tiny as 2 mm in diameter, fabricated on a silicon membrane, is a good example (see
Fig. 11.7). 15) Even this prototype motor can generate a torque three to four orders
of magnitude higher than an equivalent size silicon motor.

As the size of miniature robots/actuators decreases, the weight of the electric lead
wire connecting the power supply becomes significant, and remote control will
definitely be required for sub-millimeter devices. The photo-driven actuator
described in the previous section is a promising candidate for micro-robots.
Future of Ferroelectric Devices 299

Class lens rotor

Gold top electrode

PZT thin film Ti/Pt bottom electrode

Silicon nitride
Silicon

Fig. 11.7 Ultrasonic rotary motor as tiny as 2 mm in diameter fabricated on a

silicon membrane.

Development Pace

We need to consider a suitable research and development pace so as to introduce

new concepts and products not too early, but not too late either. Three years prior to
the commercialization is a good target for the ferroelectric device field. The Ford
company changed their development pace from 5 to 3 years several years ago, and
commercialized the "Taurus" successfully.

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 voltages. Does the reader know the available battery voltages?
The answers are 1.5, 3, 6, 12, 24 (automobile applications) and 250 V.

When the author collaborated with COPAL to develop piezoelectric camera shutters
using a bimorph structure, we initally used conventionally commercialized
bimorphs driven at around 100 V. But, when we tried to commercialize it, we
recognized that we needed an additional 100 V power supply, which would cost
more than a couple of dollars. Thus, we needed to change the bimorph design, by
thickening the layers, 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).

The reader needs to collect the necessary information on the specifications:

*sensitivity
*size
*lifetime
*available power supply
300 Chapter 11

Design

Even when the performance of the equipment is almost the same, the sales
promotion depends strongly also on the design, color, etc. The design must fit
social trends as discussed in the Marketing Creativity section.

Table 11.6 summarizes the difference in development concepts between the United
States and Japan. Does the reader remember a TV commercial about the Samsonite
suitcase? In a three-minute broadcast, a suitcase is thrown away from a tall 10-
storey building, and reaches the ground, but the suitcase is not damaged. This ad
provides the viewers with images of highly reliable and tough Samsonite suitcases.
I appreciate this typical American aptitude, but am concerned that they are not
considering glass cosmetic bottles in the suitcase.

Table 11.6 Difference in development concepts between the United States and
Japan.
___________________________________________________
U.S. Japan__________

Best device Good device

For military For civil consumers
With top technology With improved technology
At expensive price At cheap price

(Reliability) (Newness, Timing)

___________________________________________________

The promotion of Game Boy by Nintendo is based on very different concepts;

newness and timing. The game may not have sold well due simply to a
misjudgement in timing. Moreover, when Nintendo developed a "family
computer," they seemed to put a particular focus on the connector between the
software board and the hardware chassis. They seemed to try to "weaken" the
connector part to be damaged after about 100 exchanges, which corresponds to 3 - 6
month usage for a normal kid. When the toy becomes out of order and the kid
complains to his/her mother, the response from the mother (particularly Japanese)
may be, "You played TV games too much! It is about time you returned to your
studies!" In this scenario, no mother would bother to make a complaint to
Nintendo. Of course, that kid cannot endure without a family computer, once he
has learned of its fun. He may even spend his personal money to purchace another
game machine; if this is true, I admire Nintendo's strategy for their toys.

Smart Systems

"Intelligent" or "smart" materials, structures and systems are used often these days.
The bottom line of "smartness" is to possess both "sensing" and "actuating"
functions. The author describes his personal opinion on this issue as an epilogue.
Future of Ferroelectric Devices 301

When a new "sensing" function is required, most researchers are trying to add an
additional component into a system, leading to a more complex system, which is
likely to be also more bulky and expensive. The author's group is contributing
enormously to adding new functions to the conventional materials and structures,
while reducing the number of components in the system, and aiming for
miniaturization 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 Fig. 11.8, using
the case of ultrasonic motors. Starting from a propagating-wave type motor with
two piezo-actuators and two power supplies, some groups moved to a more
complex motor with 4 piezo-actuators. However, our group took the opposite
approach, simplification, and developed a standing-wave type with a single actuator
element.

Fig. 11.8 Development concept for smart systems, using as an example ultrasonic
motors.
302 Chapter 11

CHAPTER ESSENTIALS__________________________________
1. Applications of ferroelectrics --

(1) high permittivity dielectrics,

(2) ferroelectric memories,
(3) pyroelectric devices,
(4) piezoelectric devices,
(5) electrooptic devices,
(6) PTC materials, and
(7) composite materials.

2. Present market shares of ferroelectric devices -- US $2 Billion

(1) capacitors
(2) piezoelectric devices
(3) thermistors

3. Reliability issues of ferroelectric devices:

a. Reliability of ceramics --
reproducibility of ceramics, temperature characteristics, electric field
and stress dependence of properties, aging effect
b. Reliability of devices --
electrode materials, electrode designs, layer thickness dependence,
failure detection techniques
c. Drive techniques --
pulse drive method, heat generation mechanism, high power drive
technique

4. Bestseller devices --

a. Business strategy
technology, product planning, and marketing creativities
b. Marketing creativity
choose your customers,
narrow your focus and
dominate your market.
c. Technological creativity
serendipity, analogy, product effect
d. Product planning creativity
seeds and needs, developing speed, specifications

5. Directions of smart systems --

a. Adding components for higher function
b. Reducing components for miniaturization and lower cost
___________________________________________________________________
Future of Ferroelectrric Devices 303

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, MA (1996).
4) K. Uchino: Proc. 9th Int'l. Symp. Appl. Ferroelectrics, 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: Amer. Ceram. Soc. Annual Mtg. Proc., SXIX-37-96,
Indianapolis, April (1996).
9) J. Zheng, S. Takahashi, S. Yoshikawa, K. Uchino and J. W. C. de Vries: J. Amer.
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) M. Treacy and F. Wiersema: Discipline of Market Leaders, Addison-Wesley
Publishing, MA (1996).
12) Y. Hiroshima: Product Planning in the Feeling Consumer Era (1996).
13) R&D Innovator, 4, No.3, Winston J. Brill & Associates (1995).
14) Battelle Company Report (1995).
15) A. M. Flyn, L. S. Tavrow, S. F. Bart, R. A. Brooks, D. J. Ehrlich, K. R.
Udayakumar and L. E. Cross: J. Microelectro-mechanical Systems, 1, 44 (1992).