Journal J. Am. Ceram. Soc.

, 82 [4] 797–818 (1999)

Ferroelectric Ceramics: History and Technology

Gene H. Haertling*,**
Department of Ceramic and Materials Engineering, Clemson University, Clemson, South Carolina 29634-0907

Ferroelectric ceramics were born in the early 1940s with sonic transducers, radio and communication filters, pyroelec-
the discovery of the phenomenon of ferroelectricity as the tric security surveillance devices, medical diagnostic
source of the unusually high dielectric constant in ceramic transducers, stereo tweeters, buzzers, gas ignitors, positive tem-
barium titanate capacitors. Since that time, they have been perature coefficient (PTC) sensors and switches, ultrasonic mo-
the heart and soul of several multibillion dollar industries, tors, electrooptic light valves, thin-film capacitors, and ferro-
ranging from high-dielectric-constant capacitors to later electric thin-film memories.
developments in piezoelectric transducers, positive tem- The history of the discovery of ferroelectricity (electrically
perature coefficient devices, and electrooptic light valves. switchable spontaneous polarization) is a fascinating one that
Materials based on two compositional systems, barium ti- extends as far back as the mid-1600s when Rochelle salt (so-
tanate and lead zirconate titanate, have dominated the field dium potassium tartrate tetrahydrate) was first prepared by Elie
throughout their history. The more recent developments in Seignette in La Rochelle, France, for medicinal purposes.
the field of ferroelectric ceramics, such as medical ultra- However, it was approximately 200 years later before this wa-
sonic composites, high-displacement piezoelectric actuators ter-soluble, crystalline material would be investigated for its
(Moonies, RAINBOWS), photostrictors, and thin and thick pyroelectric (thermal–polar) properties, another half century
films for piezoelectric and integrated-circuit applications before its piezoelectric (stress–polar) properties would be un-
have served to keep the industry young amidst its growing covered, and finally another 40 years would pass before ferro-
maturity. Various ceramic formulations, their form (bulk, electricity (a hypothetical but yet unproved property of solids at
films), fabrication, function (properties), and future are de- the turn of the 20th century) would be first discovered by
scribed in relation to their ferroelectric nature and specific Joseph Valasek in this same material.1 Rochelle salt was a
areas of application. popular material in these initial studies, because it was readily
available and easily grown as large single crystals of excellent
optical quality, but its water solubility eventually led to its
I. Introduction
disuse in later years. Several excellent papers on the history of
ferroelectricity have been written, and the reader is referred to
S INCE the discovery of ferroelectricity in single-crystal ma-
terials (Rochelle salt) in 1921 and its subsequent extension
into the realm of polycrystalline ceramics (barium titanate,
these for many of the details.2–6
This paper is intended to cover only the highlights of ferro-
BaTiO3) during the early to mid-1940s, there has been a con- electric ceramics and cannot hope to treat all of its diverse
tinuous succession of new materials and technology develop- aspects. In this regard, only personalities and companies in-
ments that have led to a significant number of industrial and volved in the early history are specifically mentioned, although
commercial applications that can be directly credited to this it is clearly recognized that, since then, there have been many
most unusual phenomenon. Among these applications are high- excellent individuals and institutions that have been involved in
dielectric-constant capacitors, piezoelectric sonar and ultra- the research, development, and application of these very inter-
esting materials.
(1) Chronological History of Ferroelectric Materials
A chronological listing of many of the more notable specific
B. M. Kulwicki—contributing editor events in the history of ferroelectric materials is given in Table
I. Because this article emphasizes a comprehensive review of
ferroelectric (FE) polycrystalline ceramics from a materials
point of view, timeline events involving compositions, process-
Manuscript No. 189612. Received January 20, 1999; approved March 1, 1999.
Presented at the 100th Annual Meeting of The American Ceramic Society, Cin- ing, fabrication techniques, properties, patents, and applica-
cinnati, OH, May 4, 1998 (Centennial Symposium on Perspectives on Ceramic and tions are all included in Table I, whereas the specifics involv-
Glass Science and Technology, Paper No. SXVIII-007-98).
*Member, American Ceramic Society. ing ferroelectric single crystals and the development of the
**Fellow, American Ceramic Society. phenomenological basis for the ferroelectric phenomenon are

centennialfeature
797
798 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Table I. Notable Events in the History of research and development of transparent electrooptic lead lan-
Ferroelectric Materials thanum zirconate titanate (PLZT) ceramics the late 1960s, (4)
Timeline Event† the engineered ferroelectric composites of the late 1970s, (5)
the development of lead magnesium niobate (PMN) relaxor
1824 Pyroelectricity discovered in Rochelle salt
1880 Piezoelectricity discovered in Rochelle salt, quartz, ceramics and the use of sol–gel techniques for the preparation
and other minerals of ferroelectric films in the 1980s, (6) the strain-amplified ac-
tuators of the early 1990s, and (7) the current integrated fer-
1912 Ferroelectricity first proposed as property of solids roelectric films on silicon. Many of the items listed in Table I
1921 Ferroelectricity discovered in Rochelle salt are described in detail in separate sections throughout the
paper.
1935 Ferroelectricity discovered in KH2PO4
1941 BaTiO3 high-K (>1200) capacitors developed (2) Birth of Ferroelectric Ceramics
1944 Ferroelectricity discovered in ABO3-type perovskite The story of the discovery of ferroelectricity and piezoelec-
BaTiO3 tricity in ceramic materials is equally fascinating and began in
1945 BaTiO3 reported as useful piezo transducer, Pat. the early 1940s under a cloud of secrecy, because World War
No. 2 486 560 II was under way. Spurred on by the pressing need for higher-
1949 Phenomenological theory of BaTiO3 introduced dielectric-constant capacitors than could be obtained from ste-
1949 LiNbO3 and LiTaO3 reported as FE atite, mica, TiO2, MgTiO3, and CaTiO3 (K ⱕ 100), unpub-
1951 Concept of antiferroelectricity introduced lished work by Thurnauer7 and Wainer and Solomon8 firmly
1952 PZT reported as FE solid-solution system, phase established BaTiO3 as a new type of ceramic capacitor with
diagram established K > 1100. Near the end of World War II, in the mid-1940s,
1953 PbNb2O6 reported as FE publications began to appear in the open literature, and it be-
1954 PZT reported as useful piezo transducer, Pat. No. came evident that concurrent work on BaTiO3 as a high-
2 708 244 dielectric-constant material had been conducted by several
1955 PTC effect in BaTiO3 reported countries, including the United States, United Kingdom,
1955 Chemical coprecipitation of FE materials
introduced USSR, and Japan. Shortly thereafter, in 1945 and 1946, the
1955 Alkali niobates reported as FE work of Wul and Goldman9 in the USSR and von Hippel’s
1957 BaTiO3 barrier layer capacitors developed group10 at the Massachusetts Institute of Technology estab-
1959 PZT 5A and 5H MPB-type piezo compositions, lished that the source of the high dielectric constant in BaTiO3
Pat. No. 2 911 370 emanated from its ferroelectric properties. Work on single-
1961 Lattice dynamics theory for FE materials, soft
crystal BaTiO3 subsequently corroborated these findings.
modes introduced The knowledge of the ferroelectric nature of ceramic BaTiO3
1961 PMN relaxor materials reported proved to be invaluable when it was discovered by Gray11 (in
1964 Oxygen/atmosphere sintering for FEs developed 1945) that an external electric field could orient the domains
1964 FE semiconductor (PTC) devices developed within the grains, thus producing a ceramic material that acted
1967 Optical and E/O properties of hot-pressed FE very similar to a single crystal possessing both ferroelectric and
ceramics reported piezoelectric properties. This electrical aligning, or “poling”
1969 Terms “ferroic” and “ferroelasticity” introduced process as it has come to be called, was thus correctly identi-
1969 Optical transparency achieved in hot-pressed PLZT fied as the key to turning an inert ceramic into an electrome-
ceramics chanically active material with a multitude of industrial and
1970 PLZT compositional phase diagram established, commercial uses. This was a most startling discovery, because
Pat. No. 3 666 666 the prevailing opinion was that ceramics could not be piezo-
1971 Useful E/O properties reported for PLZT, Pat. No. electrically active, because the sintered and randomly oriented
3 737 211 crystallites would, on the whole, cancel out each other. This
1973 Oxygen/atmosphere sintering of PLZT to full
transparency
1977 FE thin films developed
1978 Engineered (connectivity designed) FE composites
developed
Abbreviations Used
1980 Electrostrictive relaxor PMN devices developed, Ferroelectric Materials
Pat. No. 5 345 139
1981 Sol–gel techniques developed for the preparation of PZT Lead zirconate titanate
FE films
1983 Photostrictive effects reported in PZT and PLZT PLZT Lead lanthanum zirconate titanate
PMN Lead magnesium niobate
1991 Moonie piezo flextensional devices developed, Pat. PT Lead titanate
No. 4 999 819 PZN Lead zinc niobate
1992 RAINBOW piezo bending actuators developed, Pat. PSZT Lead stannate zirconate titanate
No. 5 471 721
1993 Integration of FE films to silicon technology, Pat. PZ Lead zirconate
No. 5 038 323 BST Barium strontium titanate
1997 Relaxor single-crystal materials developed for piezo SBT Strontium bismuth titanate
transducers
†
Others
FE is ferroelectric, K is relative dielectric constant, PZT is lead zirconate titanate,
MPB is morphotropic phase boundary, PLZT is lead lanthanum zirconate titanate, FE Ferroelectric
PMN is lead magnesium niobate, PTC is positive temperature coefficient, E/O is
electrooptic, RAINBOW is reduced and internally biased oxide wafer. AFE Antiferroelectric
PE Paraelectric
SFE Slim-loop ferroelectric
PTC Positive temperature coefficient
not treated in detail. The time period is from the early 1800s to NTC Negative temperature coefficient
the present (1999), involving events from the early work on MLC Multilayer capacitor
single-crystal Rochelle salt to (1) the birth of ferroelectric ce- BLC Barrier layer capacitor
ramics in the 1940s, (2) the development of lead zirconate MPB Morphotrophic phase boundary
titanate (PZT) piezoelectric ceramics in the mid-1950s, (3) the
April 1999 Ferroelectric Ceramics: History and Technology 799

proved not to be the case for ferroelectric crystallites, because

they could be permanently aligned or reoriented in an electric
field, somewhat analogous to magnetic alignment in permanent
magnets.
Thus, as pointed out by Jaffe12 in his excellent treatise on
piezoelectric ceramics, the three fundamental steps that were
critical to the understanding of ferroelectricity and piezo-
electricity in ceramics were (1) the discovery of the unusually
high dielectric constant of BaTiO3, (2) the discovery that the
origin of the high dielectric constant was due to its ferroelectric
(permanent internal dipole moment) nature, thus ushering in a
new class of ferroelectrics (the simple oxygen octahedral
ABO3 group), and (3) the discovery of the electrical poling
process that aligns the internal dipoles of the crystallites
(domains) within the ceramic and causes it to act very simi-
lar to a single crystal. For more details on the history of fer-
roelectric ceramics, the reader is referred to several excellent
publications.13–15
(3) Basis for Piezoelectricity in Solids
Piezoelectricity, a property possessed by a select group of
materials, was discovered in 1880 by Jacques and Pierre Curie
during their systematic study of the effect of pressure on the
generation of electrical charge by crystals, such as quartz,
zincblende, and tourmaline. The name “piezo” is derived from
the Greek, meaning “to press;” hence, piezoelectricity is the
generation of electricity as a result of a mechanical pressure.
Cady16 defines piezoelectricity as “electric polarization pro-
duced by mechanical strain in crystals belonging to certain
classes, the polarization being proportional to the strain and
changing sign with it.”
An understanding of the concept of piezoelectricity in solids Fig. 1. Interrelationship of piezoelectric and subgroups on the basis
begins with an understanding of the internal structure of the of symmetry.
material; for purposes here, consider a single crystallite. This
crystallite has a definite chemical composition and, hence, is
made up of ions (atoms with positive or negative charge) that (4) Piezoelectricity in Ferroelectric Ceramics
are constrained to occupy positions in a specific repeating re- As mentioned previously, the poling process is the critical
lationship to each other, thus building up the structure or lattice element in being able to utilize the piezoelectric effect in a
of the crystal. The smallest repeating unit of the lattice is called ferroelectric ceramic. Without poling, the ceramic is inactive,
the unit cell, and the specific symmetry possessed by the unit even though each one of the individual crystallites is piezo-
cell determines whether it is possible for piezoelectricity to electric itself. With poling, however, the ceramic becomes ex-
exist in the crystal. Furthermore, the symmetry of a crystal’s tremely useful, provided that it is not heated above its Curie
internal structure is reflected in the symmetry of its external temperature (TC), where it loses its polarization and all of
properties (Neumann’s principle).16 the orientation of the polarization produced by the poling
The elements of symmetry that are utilized by crystallogra- process.17
phers to define symmetry about a point in space, e.g., the Two effects are operative in piezoelectric crystals, in gen-
central point of a unit cell, are (1) a center of symmetry, (2) eral, and in ferroelectric ceramics, in particular. The direct
axes of rotation, (3) mirror planes, and (4) combinations of effect (designated as a generator) is identified with the phe-
these. All crystals can be divided into 32 different classes or nomenon whereby electrical charge (polarization) is generated
point groups utilizing these symmetry elements, as shown in from a mechanical stress, whereas the converse effect (desig-
Fig. 1. These 32 point groups are subdivisions of seven basic nated as a motor) is associated with the mechanical movement
crystal systems that are, in order of ascending symmetry, tri- generated by the application of an electrical field. Both of these
clinic, monoclinic, orthorhombic, tetragonal, rhombohedral effects are illustrated in Fig. 2 as cartoons for easy grasp of the
(trigonal), hexagonal, and cubic. Of the 32 point groups, 21 principles.
classes are noncentrosymmetric (a necessary condition for pi- The basic equations that describe these two effects in regard
ezoelectricity to exist) and 20 of these are piezoelectric. One to electric and elastic properties are13
class, although lacking a center of symmetry, is not piezoelec-
tric because of other combined symmetry elements. A lack of D ⳱ dE + ␧TE (generator) (1)
a center of symmetry is all-important for the presence of pi- S ⳱ s T + dE
E
(motor) (2)
ezoelectricity when one considers that a homogeneous stress is
centrosymmetric and cannot produce an unsymmetric result, where D is the dielectric displacement (consider it equal to
such as a vector-quantity-like polarization, unless the material polarization), T the stress, E the electric field, S the strain, d a
lacks a center of symmetry, whereby a net movement of the piezoelectric coefficient, s the material compliance (inverse of
positive and negative ions with respect to each other (as a result modulus of elasticity), and ␧ the dielectric constant (permittiv-
of the stress) produces electric dipoles, i.e., polarization. Fur- ity). The superscripts indicate a quantity held constant: in the
thermore, for those materials that are piezoelectric but not fer- case of ␧T, the stress is held constant, which means that the
roelectric (i.e., they do not possess spontaneous polarization), piezoelectric element is mechanically unconstrained, and, in
the stress itself is the only means by which the dipoles are the case of sE, the electric field is held constant, which means
generated. For piezoelectricity, the effect is linear and revers- the electrodes on the element are shorted together. Equations
ible, and the magnitude of the polarization is dependent on the (1) and (2), in matrix form, actually describe a set of equations
magnitude of the stress and the sign of the charge produced is that relate these properties along different orientations of the
dependent on the type of stress (tensile or compressive). material. Because of the detailed nature of the many equations
800 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Fig. 2. Piezoelectric effects in ferroelectric ceramics.

involved, the reader is referred to several sources on the sub- ergy (or vice versa) is always incomplete, k is always less than
ject.13,18–20 Suffice it to say that, because this is a piezoelectric unity. Commonly used as a figure-of-merit for piezoelectrics,
solid, Eqs. (1) and (2) relate given properties, such as electric the higher k values are most desirable and constantly sought
displacement (polarization) and strain to both the mechanical after in new materials. For ceramics, kp is a typical measure-
and electrical states of the material. Furthermore, these prop- ment used in comparing materials–values ranging from 0.35
erties are directional quantities, and, hence, they are usually for BaTiO3 to as high as 0.72 for PLZT.
specified with subscripts to identify the conditions under which All of the properties mentioned here may be realized in a
they are determined, e.g., d31 indicates that this piezoelectric piezoelectric ceramic, which is, in reality, a poled ferroelectric
coefficient relates to the generation of polarization (direct ef- ceramic material. During the process of poling, there is a small
fect) in the electrodes perpendicular to the 3 or vertical direc- expansion of the material along the poling axis and a slight
tion and to the stress mechanically applied in the 1 or lateral contraction in both directions perpendicular to it. The strength
direction; d33 indicates the polarization generated in the 3 di- of the poling field, often in combination with elevated tem-
rection when the stress is applied in the 3 direction. Typical perature, is an important factor in determining the extent of
relationships for this coefficient are: alignment and, hence, the resulting properties. Alignment is
never complete; however, depending on the type of crystal
D3 ⳱ d33T3 (direct effect) (3) structure involved, the thoroughness of poling can be quite
S3 ⳱ d33E3 (converse effect) (4) high, ranging from 83% for the tetragonal phase to 86% for the
rhombohedral phase, and to 91% for the orthorhombic phase,
where the d coefficients are numerically equal in both equa- when compared with single-domain, single-crystal values. Be-
tions. The d coefficients are usually expressed as ×10−12 C/N cause all ceramic bodies are macroscopically isotropic in the
for the direct effect and ×10−12 m/V for the converse effect. “as-sintered” condition and must be poled to render them use-
High d coefficients are desirable for those materials that are ful as piezoelectric materials, they are all ferroelectric as well
utilized in motional or vibrational devices, such as sonar and as piezoelectric.
sounders. (5) Basis for Ferroelectricity in Ceramics
In addition to the d coefficients, open-circuit g coefficients
are also used to evaluate piezoelectric ceramics for their ability Figure 1 shows that there are 10 crystal classes out of a
to generate large amounts of voltage per unit of input stress. possible 20 that are designated as pyroelectric. This group of
The g constant is related to the d constant via the relationship materials possesses the unusual characteristic of being perma-
nently polarized within a given temperature range. Unlike the
d more general piezoelectric classes that produce a polarization
g= (5) under stress, the pyroelectrics develop this polarization spon-
K␧0
taneously and form permanent dipoles in the structure. This
where K is the relative dielectric constant and ␧0 the permit- polarization also changes with temperature—hence, the term
tivity of free space (8.854 × 10−12 F/m). Thus, a high g constant pyroelectricity. Pyroelectric crystals, such as tourmaline and
is possible for a given d coefficient if the material has a low K. wurtzite, are often called polar materials, thus referring to the
High-g-constant ceramics are usually ferroelectrically hard ma- unique polar axis existing within the lattice. The total dipole
terials that do not switch their polarization readily and possess moment varies with temperature, leading to a change in sign
lower K values. They are used in devices such as portable gas for the current flowing out of a short-circuited crystal.
ignitors and patio lighters. A subgroup of the spontaneously polarized pyroelectrics is a
The piezoelectric coupling factor (e.g., k33, k31, and kp) is a very special category of materials known as ferroelectrics.
convenient and direct measurement of the overall strength of Similar to pyroelectrics, materials in this group possess spon-
the electromechanical effect, i.e., the ability of the ceramic taneous dipoles; however, unlike pyroelectrics, these dipoles
transducer to convert one form of energy to another. It is de- are reversible by an electric field of some magnitude less than
fined as the square root of the ratio of energy output in elec- the dielectric breakdown of the material itself. Thus, the two
trical form to the total mechanical energy input (direct effect), conditions necessary in a material to classify it as a ferroelec-
or the square root of the ratio of the energy available in me- tric are (1) the existence of spontaneous polarization and (2) a
chanical form to the total electrical energy input (converse demonstrated reorienting of the polarization.
effect). Because the conversion of electrical to mechanical en- Four types of ceramic ferroelectrics are also given in Fig. 1
April 1999 Ferroelectric Ceramics: History and Technology 801

as subcategories of the general group of ferroelectric materials, a tetragonal structure, although it should be understood that it
with typical examples representing the type based on its unit- can also occur along the orthogonal a or b axes as well. The
cell structure: (1) the tungsten–bronze group, (2) the oxygen views of “polarization up” and “polarization down” (represent-
octahedral group, (3) the pyrochlore group, and (4) the bismuth ing 180° polarization reversal) show two of the six possible
layer–structure group. Of these, the second group (ABO3 permanent polarization positions.
perovskite type) is by far the most important category, eco- When many of these unit cells, which are adjacent to each
nomically. The families of compositions listed (BaTiO3, PZT, other, switch in like manner, this is referred to as domain
PLZT, PT (lead titanate), PMN, and (Na,K)NbO3) represent reorientation or switching. The homogeneous areas of the ma-
the bulk of the ferroelectric ceramics manufactured in the terial with the same polarization orientation are referred to as
world today. domains, with domain walls existing between areas of unlike
A typical ABO3 unit-cell structure is given in Fig. 3. For polarization orientation. There exists in tetragonal materials
example, the PLZT unit cell consists of a corner-linked net- both 90° (strain-producing domains on switching) and 180°
work of oxygen octahedra with Zr4+ and Ti4+ ions occupying domains (nonstrain-producing domains), whereas the strain-
sites (B sites) within the octahedral cage and the Pb2+ and producing entities in rhombohedral materials are 71° and 109°
La3+ ions situated in the interstices (A sites) created by the domains with the 180° domains remaining as nonstrain pro-
linked octrahedra. As a result of the different valency between ducing. Macroscopic changes occur in the dimensions of the
Pb2+ and La3+, some of the A sites and B sites are vacant material when strain-producing domains are switched.
(referred to as vacancies) to maintain electrical neutrality in the Because of the empirical nature of determining the revers-
structure. ibility of the dipoles (as detected by a hysteresis loop measure-
When an electric field is applied to this unit cell, the Ti4+ or ment), one cannot predict the existence of ferroelectricity in a
Zr4+ ion moves to a new position along the direction of the new material with much accuracy. However, the basis for the
applied field. Because the crystallite and, hence, the unit cell is existence of ferroelectricity rests primarily on structural and
randomly oriented and the ions are constrained to move only symmetry considerations. The special relationship of ferroelec-
along certain crystallographic directions of the unit cell, it is trics as a subgroup of piezoelectrics (Fig. 1) infers that “all
most often the case that an individual ionic movement only ferroelectrics (poled) are piezoelectric, but not all piezoelec-
closely approximates an alignment with the electric field. How- trics are ferroelectric.” The current number of ferroelectrics is
ever, when this ionic movement does occur, it leads to a mac- in the thousands when one includes the many ceramic solid-
roscopic change in the dimensions of the unit cell and the solution compositions. Ferroelectrics are no longer the great
ceramic as a whole. The dimensional change can be as large as “accident of nature” that they were once thought to be in the
a few tenths of a percent elongation in the direction of the field 1920s and 1930s.
and approximately one-half that amount in the other two or-
thogonal directions. The original random orientation of the (6) Electrostriction in Ferroelectric Ceramics
domain polarization vectors (virgin condition) can be restored Electrostriction is another electromechanical effect that ex-
by heating the material above its TC. This process is known as ists in ferroelectric ceramics. In electrostriction, the sign of the
thermal depoling. deformation that occurs with an electric field is independent of
Also shown in Fig. 3 is the reversibility of the polarization the polarity of the field and is proportional to even powers of
caused by the displacement of the central Ti4+ or Zr4+ ion. the field. In piezoelectricity, the deformation is linear with
Displacement is illustrated here as occurring along the c axis in respect to the applied field and changes sign when the field is
reversed. This means in practical terms that electrostriction
produces an expansion in most materials in the direction of the
field regardless of its polarity, and this expansion relaxes back
to zero when the field is removed. The corresponding equations
are
S ⳱ mE2 (in terms of electric field) (6)
S ⳱ QP2 (in terms of polarization) (7)
where P is the polarization and m and Q the corresponding
electrostrictive coefficients. Similar to piezoelectricity, elec-
trostriction deals with vector quantities, and, hence, appropriate
subscripts must be used.
Although electrostriction is a general property of all dielec-
tric materials, whether they are crystalline, amorphous, polar,
or centrosymmetric, it can be particularly large in ferroelectric
materials just above their TC, where an electric field can en-
force the energetically unstable ferroelectric phase. More com-
monly, this effect is utilized to good advantage in relaxor ma-
terials, such as PMN, PZN (lead zinc niobate), and PLZT,
where the TC is not sharp but rather is spread out over a mod-
erate temperature range, thus allowing for a reasonable tem-
perature range of operation for devices made from them.
Electrostrictive materials can be operated either in the elec-
trostrictive mode (as stated above) or in the field-biased piezo-
electric mode. In the latter case, a dc electric field bias is
applied to the material to induce a ferroelectric polarization,
whereupon the material acts as a normal piezoelectric as long
as the field is applied, and, the stronger the field, the higher the
piezoelectric effect until saturation sets in. The relationships
relating the resulting piezoelectric d coefficient to the induced
Fig. 3. Perovskite ABO3 unit cell for PZT or PLZT, illustrating 180° polarization and the dielectric permittivity are
polarization reversal for two of the six possible polarization states
produced by displacement of the central cation in the tetragonal plane. d33 ⳱ 2Q11P3␧33 (8)
802 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

d31 ⳱ 2Q12P3␧33 (9) (7) Electrooptic Effects

The electrooptic properties of PLZT materials are intimately
where Q11 and Q12 are the longitudinal and transverse elec- related to their ferroelectric properties. Consequently, varying
trostrictive coefficients, respectively. Thus, a large dielectric the ferroelectric polarization with an electric field, such as in a
constant and a high polarization are required to produce a large hysteresis loop, also produces a change in the optical properties
induced piezoelectric coefficient. These Q coefficients are es- of the ceramic. Moreover, the magnitude of the observed elec-
sentially independent of temperature.21–24 trooptic effect is dependent on both the strength and direction
Some advantages for electrostrictors over conventional pi- of the electric field.
ezoelectrics are (1) hysteresis in the strain-field dependence is PLZT ceramics display optically uniaxial properties on a
minimal or negligible in a selected temperature range, (2) re- microscopic scale, and also on a macroscopic scale when po-
alized deformation is more stable and comparable to the best larized or activated with an electric field. There is one unique
piezoelectric ceramics, and (3) no poling is required. These symmetry axis in uniaxial crystals, the optic axis (colinear with
advantages, however, are balanced by the disadvantages of (1) the ferroelectric polarization vector in ceramic PLZT), which
a limited usable temperature range due to the strong tempera- possesses optical properties different from the other two or-
ture dependence of the electrostrictive effect and (2) especially thogonal axes. That is, light traveling in a direction along the
small deformation at low fields, because electrostriction is very optical axis and vibrating in a direction perpendicular to it
nearly a quadratic function at low electric fields, which then possesses an index of refraction (n0) different from light trav-
usually necessitates higher operating voltages to achieve mod- eling in a direction 90° to the optic axis and vibrating parallel
erate deflection. to it (ne). The absolute difference between the two indexes is
Although the actual mechanism or mechanisms leading to defined as the birefringence; i.e., ne − n0 ⳱ ⌬n. On a macro-
the large electrostrictive effects is not fully understood, it is scopic scale, ⌬n is equal to zero before poling and has some
generally believed that the electrostrictive effect in these ma- finite value after poling, depending on the composition and the
terials is due to the field-activated coalescence of micropolar degree of poling. With relaxor materials, ⌬n is not permanent
regions to macrodomains of the parent ferroelectric.23,25 Con- but exists only as long as the electric field is present.27
sequently, the mechanism is essentially the same as the non- A typical setup for determining the behavior of electrooptic
polar, cubic prototype of the ferroelectric phase undergoing a ceramics is given in Fig. 4. Linearly polarized white light, on
momentary phase transformation to the ferroelectric phase entering the electrically energized ceramic, is split in two or-
while under the influence of an electric field. Furthermore, thogonally vibrating white light components (represented in
because a higher applied electric field leads to a larger mag- Fig. 4 by red and green waves), whose vibration directions are
nitude of the induced ferroelectric polarization and strain, one defined by the crystallographic axes of the crystallites acting as
is able to use this effect for achieving voltage-dependent di- one optical entity; in this case, the axes are defined by the
electric, piezoelectric, pyroelectric, and electrooptic properties direction of the electric field. Because of the different refrac-
in ceramics. Longitudinal strains as high as 0.1% in PMN and tive indices, ne and n0, the propagation velocity of the two
0.3% for PLZT (La/Zr/Ti ⳱ 9/65/35) have been reported for components is different within the material and results in a
these electrostrictive materials. phase shift known as optical retardation. The total retardation
Electrostriction-like effects are also evident in nonpolar, an- (⌫) is a function of both ⌬n and the optical path length (t),
tiferroelectric (antipolar adjacent unit cells) materials that un- according to the relationship
dergo a phase change from antiferroelectric (AFE) to ferroelec-
tric when a sufficiently high electric field is applied. Although ⌫ ⳱ ⌬nt (10)
this effect is very abrupt (occurring suddenly at a specific volt-
age) in a material such as lead stannate zirconate titanate where t is the ceramic thickness along the optical path. In the
(PSZT), in another ceramic, such as PLZT, the change is case of white light (average wavelength of 0.55 ␮m), as shown
“washed out” over a range of voltages; i.e., the former is more in Fig. 4, when sufficient voltage is applied to the ceramic, a
digital, and the latter is more analog. Longitudinal strains as half-wave retardation is achieved for one component relative to
high as 0.8% have been reported in these antiferroelectric the other. The net result is one of rotating the vibration direc-
materials.26 tion of the polarized light by 90°, thus allowing it to be trans-

Fig. 4. Basic setup for evaluating electrooptic shutter/modulator characteristics (open condition shown).
April 1999 Ferroelectric Ceramics: History and Technology 803

mitted by the second (crossed) polarizer. Switching of the ce- or depress the sharpness of the dielectric constant peak at the
ramic from a state of zero retardation (no voltage) to half-wave TC, thus giving a flatter dielectric constant–temperature profile.
(full voltage) creates an ON/OFF light shutter, whereas selec- The net results of these efforts is to produce ceramic capacitors
tion of intermediate voltages create an analog modulator. Color with dielectric constants up to 3000, loss tangents of ∼1% or
generation (yellow, red, blue, and green) from the incoming less, and temperature stabilities of ±15% for the X7R-type
white light can be achieved by increasing the voltage beyond capacitors. Higher dielectric constants (to 12 000) can be
half-wavelength; however, if monochromatic light is used, achieved with a concurrent loss in temperature stability
extending the voltage beyond half-wavelength results only (+22%/−56%) for the Z5U-type capacitor, as designated by the
in a progression of repeating light and dark bands, as in an Electronic Industries Association (EIA).28
interferometer. Dielectric constants in the range of 100 000 have been
There are two common types of electrooptic birefringent achieved with BaTiO3-based compositions that also contain
effects within the PLZT compositional phase diagram, i.e., (1) special additives to suppress the ferroelectric properties and
nonmemory quadratic (Kerr effect) and (2) memory linear facilitate the development of a chemically reduced material
(Pockel effect). The respective electrooptic coefficients for with semiconducting properties. These are the barrier-layer ca-
these effects are calculated from the following relationships: pacitors (BLCs).29 These BLCs are produced by carefully
reoxidizng a thin barrier layer in the boundary between each of
2⌫ the individual semiconducting grains of the ceramic, and it is
R=− (11)
n tE2
3 these many insulating boundary layers that actually make up
the capacitor. Because these barrier-layer thicknesses are mea-
2⌫ sured in at ∼1–2 ␮m, this type of capacitor is limited to <50 V.
rc = − (12) Another type of material that was developed as early as 1955
n3tE
from a BaTiO3 base is the PTC ceramic possessing electrically
where R is the quadratic coefficient, rc the linear coefficient, conducting properties at room temperature and rather abruptly
and n the index of refraction (n ⳱ 2.5 for PLZT). If a polished, changing to a highly resistive material at some elevated tem-
thin plate or a film-on-substrate is to be evaluated rather than perature at TC ≈ 120°C.30,31 Changing the TC with appropriate
a cube of bulk material, interdigital electrodes are often applied additives, as mentioned previously for capacitors, changes the
to the free surface of the material as opposed to the completely temperature at which this PTC resistivity anomaly occurs. This
transverse electrodes shown in Fig. 4; in this case, the coeffi- effect is exactly opposite to the more-common effect in the
cients obtained are only a close approximation to the true co- negative temperature coefficient (NTC) materials, which expe-
efficients, because the electric field lines are not truly trans- rience a reduction in resistance on increasing temperature. In
verse, but, rather, they are constrained to penetrate the material the case of the PTCs, a small (0.2 mol%) addition of an off-
in a nonlinear manner. valent additive, such as Y3+ or La3+, is used to produce an
electrically semiconducting body without totally destroying the
ferroelectric properties of the material, even though one is not
II. Materials able to ascertain ferroelectricity because of its conductivity. In
fact, it is believed that the spontaneous polarization developed
(1) Barium Titanate Ceramics at the TC nullifies or lowers the height of the barrier at the
BaTiO3 is the first piezoelectric transducer ceramic ever de- boundary of the grains, thereby allowing easy passage of cur-
veloped; however, its use in recent years has shifted away from rent at temperatures below TC. When the temperature is in-
transducers to an almost exclusive use as high-dielectric- creased through the TC, spontaneous polarization disappears,
constant capacitors of the discrete and multilayer (MLC) types. and the barrier height is again raised, leading to an increased
The reasons for this are primarily twofold: (1) its relatively low resistance of ∼6 or 7 orders of magnitude. The barriers on the
TC of 120°C, which limits its use as high-power transducers, surface of the grains (grain boundaries) are produced by sin-
and (2) its low electromechanical coupling factor in compari- tering and cooling the material on a rigid schedule to produce
son to PZT (0.35 vs 0.65), which limits its operational output. a controlled oxidized layer. Applications include switches, sen-
Unlike PZT, which is a solid-solution composition containing sors, motor starters, and controllers. Incidentally, the PTC ef-
a volatile component (PbO), BaTiO3 is a definite chemical fect is one of the few examples where a ceramic property in a
compound possessing relatively-high-stability components, material surpasses that of the corresponding single crystal, be-
making it easy to sinter while maintaining good chemical stoi- cause the effect is absent in single crystals because there are no
chiometry. Nevertheless, these materials are not actually used grain boundaries.
in their true chemical form, but, rather, are combined with
special additives to modify and improve their basic properties. (2) PZT and PLZT Ceramics
The additives for BaTiO3 transducers usually are Sr2+ for vary- Ferroelectric ceramics for piezoelectric applications histori-
ing the TC downward from 120°C, Pb2+ for varying the TC cally have been formulated from a number of compositions and
upward, Ca2+ for increasing the temperature range of stabil- solid solutions including BaTiO3, PZT, PLZT, PbN2O6,
ity of the tetragonal phase, and Co2+ for decreasing the NaNbO3, and PT. Foremost of these has been BaTiO3, which
high-electric-field losses without affecting the piezoelectric dates from the early 1940s, but, in the past several decades, it
constants. largely has been supplanted by the PZTs and PLZTs for trans-
When BaTiO3 is used in its primary application as a capaci- ducer applications.13,32,33 This is because PZT and PLZT com-
tor, a different group of additives is used, because the intent in positions (1) possess higher electromechanical coupling coef-
this case is to suppress the ferroelectric and piezoelectric prop- ficients than BaTiO3, (2) have higher TC values, which permit
erties as much as possible while maintaining or increasing its higher temperatures of operation or higher temperatures of pro-
dielectric constant. Two general types of modifiers are com- cessing during the fabrication of devices, (3) can be easily
monly used: TC shifters and TC depressors. The TC shifters, poled, (4) possess a wide range of dielectric constants, (5) are
such as SrTiO3, CaZrO3, PbTiO3, and BaSnO3, have the effect relatively easy to sinter at lower temperatures than BaTiO3, and
of shifting TC to a higher or lower value, depending of the (6) form solid-solution compositions with many different con-
intended result. However, it is usually the case that a lower TC stituents, thus allowing a wide range of achievable properties.
is desired, such that the higher permittivity values associated PZT ceramics are almost always used with a dopant, a modi-
with the TC occur nearer room temperature or the temperature fier, or other chemical constituent or constituents to improve
of operation. Depressors, such as Bi 2 (SnO 2 ) 3 , MgZrO 3 , and optimize their basic properties for specific applica-
CaTiO3, NiSnO3, as well as the shifters, are added in small tions.13,19,34 Examples of these additives include off-valent do-
(1–8 wt%) quantities to the base BaTiO3 composition to lower nors, such as Nb5+ replacing Zr4+ or La3+ replacing Pb2+, to
804 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

counteract the natural p-type conductivity of the PZT and, thus, all compositional aspects of the dielectric, piezoelectric, pyro-
increase the electrical resistivity of the materials by at least 3 electric, ferroelectric, and electrooptic ceramics is the PLZT
orders of magnitude. The donors are usually compensated by system.35 Figure 5 shows the PLZT system with the parent PZT
A-site vacancies. These additives (and vacancies) enhance do- phase diagram. Several areas on the diagram are color coded
main reorientation; ceramics produced with these additives are for easy identification: (1) the ferroelectric tetragonal and
characterized by square hysteresis loops, low coercive fields, rhombohedral phases are shown in orange, (2) the orthorhom-
high remanent polarization, high dielectric constants, maxi- bic antiferroelectric phase in purple, (3) the cubic paraelectric
mum coupling factors, higher dielectric loss, high mechanical (PE; nonferroelectric) phases in white, (4) the morphotropic
compliance, and reduced aging. phase boundary (MPB) in magenta, (5) the pyroelectric appli-
Off-valent acceptors, such as Fe3+ replacing Zr4+ or Ti4+, are cation areas near PbTiO3 in blue, (6) the economically impor-
compensated by oxygen vacancies and usually have only lim- tant MPB compositions that embrace almost all of the trans-
ited solubility in the lattice. Domain reorientation is limited, ducer applications in green, (7) the compositional area for
and, hence, ceramics with acceptor additives are characterized AFE-to-FE, enforced-phase devices in gray, and (8) specific
by poorly developed hysteresis loops, lower dielectric con- compositions in these regions in yellow.
stants, low dielectric losses, low compliances, and higher aging Figure 5 shows that the effect of adding lanthanum to the
rates. PZT system is (1) one of maintaining extensive solid solu-
Isovalent additives, such as Ba2+ or Sr2+ replacing Pb2+ or tion throughout the system and (2) one of decreasing the sta-
Sn4+ replacing Zr4+ or Ti4+, in which the substituting ion is of bility of the ferroelectric phases in favor of the paraelectric and
the same valency and approximately the same size as the re- antiferroelectric phases, as indicated by the red line, which
placed ion, usually produce inhibited domain reorientation and shows the reduction of the TC with increasing lanthanum. At a
poorly developed hysteresis loops. Other properties include 65/35 ratio of PZ/PT (where PZ is lead zirconate, PbZrO3), a
lower dielectric loss, low compliance, and higher aging rates. concentration of 9.0% lanthanum (designated as 9/65/35) is
Dopants are usually added in concentrations of ⱕ3 at.%. sufficient to reduce the temperature of the stable ferroelectric
Modifiers are substituted into the original PZT composition as polarization to slightly below room temperature, resulting in a
solid-solution constituents in concentrations of ⱖ5 at.%. The material that is nonferroelectric and cubic in its virgin state.
most common examples of modifier systems are (Pb,La) The cross-hatched area existing along the FE–PE phase bound-
(Zr,Ti)O 3 , (Pb,Sr)(Zr,Ti)O 3 , (Pb,Ba)(Zr,Ti)O 3 , Pb- ary denotes a region of diffuse, metastable relaxor phases that
(Zr,Ti,Sn)O 3 , (Pb,La)TiO 3 , and Pb(Mg,Nb)O 3 –PbZrO 3 – can be electrically induced to a ferroelectric phase. Materials
PbTiO3, although, in actuality, there are many of these lead- within this region exhibit a quadratic strain and electrooptic
containing, solid-solution systems.13 One system that embraces behavior.

Fig. 5. Phase diagram of the PZT and PLZT solid-solution systems.

April 1999 Ferroelectric Ceramics: History and Technology 805

The solubility of lanthanum in the PZT lattice is a function equivalent to TC for normal ferroelectrics) of PMN to ∼40°C.
of composition and is related directly to the amount of PT For this composition, the temperature of polarization loss (Td)
present. The compositional dependence of the solubility limit is is ∼10°C; hence, the material is a relaxor at room temperature
indicated by the dashed line adjacent to the mixed-phase region (25°C). An addition of ∼28% PT causes the material to revert
(double cross-hatched area) in Fig. 5. For the two end-member to a normal ferroelectric tetragonal phase with TC ≈ 130°C.
compositions, PZ and PT, these limits are 4 and 32 at.%, re- Unlike PZT and PLZT, PMN ceramics are somewhat diffi-
spectively. The solubility limits for intermediate compositions cult to prepare in a phase-pure condition. Several methods of
are proportional to their Zr/Ti ratios. powder preparation have been developed over the years to
Modification of the PZT system by the addition of lantha- reduce the undesirable pyrochlore phase to a bare minimum,41
num sesquioxide has a marked beneficial effect on several of but the process that has met with consistent success is the
the basic properties of the material, such as increased square- so-called columbite precursor method.42 In this technique,
ness of the hysteresis loop, decreased coercive field, increased MgO and Nb2O5 are first reacted to form the columbite struc-
dielectric constant, maximum coupling coefficients, increased ture, MgNb2O6, which is then reacted with PbO and TiO2 to
mechanical compliance, and enhanced optical transparency. form the PMN–PT compositions.
The optical transparency was discovered in the late 1960s as a
result of an in-depth study of various additives to the PZT (4) Ferroelectric Films
system.35 Results from this work indicate that La3+, as a chemi- The 1970s and 1980s witnessed the emergence of thin and
cal modifier, is unique among the off-valent additives in pro- thick films (both ferroelectric and nonferroelectric) as an im-
ducing transparency. The reason for this behavior is not fully portant category of materials that was brought about by the
understood; however, it is known that lanthanum is, to a large maturing of laser and transistor technologies (e.g., optical fi-
extent, effective because of its high solubility in the oxygen bers, integrated optics, microelectromechanical systems, mi-
octrahedral structure, thus producing an extensive series of croprocessors, and computers) and promises to be the spring-
single-phase, solid-solution compositions. The mechanism is board for the age of integration beyond the 1990s into the next
believed to be one of lowering the distortion of the unit cell, century.43,44 New materials development during this time pe-
thereby reducing the optical anisotropy of the unit cell and, at riod was one of form (i.e., from bulk to films) rather than
the same time, promoting uniform grain growth and densifica- composition. Almost all of the current compositions that are
tion of a single-phase, pore-free microstructure. used in the fabrication of films had their beginnings in the bulk
Electrooptic compositions in the PLZT phase diagram are materials. Examples of these include BaTiO3, barium strontium
generally divided into three application areas: (1) nonmemory titanate, PZT, PLZT, PNZT(Nb), PSZT(Sn), PBZT(Ba), PT,
quadratic, (2) memory, and (3) linear. As mentioned previ- bismuth titanate, lithium niobate, barium strontium niobate,
ously, the quadratic materials are located along the FE–PE strontium barium tantalate, and potassium niobate. Thus, one
phase boundary, principally in the cross-hatched area. Memory can say (at least for the present time) that an adequate number
compositions having stable, electrically switchable polarization of ferroelectric compositions now exist and are being produced
and optical states are largely located in the ferroelectric rhom- as good-quality, polycrystalline thin and thick films by a vari-
bohedral phase region, and the linear materials possessing non- ety of forming methods. These films will form the basis for the
switching, linear strain, and electrooptic effects are confined to development of new structures and devices well beyond the
the area encompassing the tetragonal phase. turn of the century.
(3) PMN Ceramics
Although the study of relaxor materials began in the early III. Processing
1960s with work on single-crystal Pb(Mg1/3Nb2/3)O3 (PMN)
materials36 and continued in the mid-1960s with PMN as one (1) Powder Preparation
of the triaxial components in the PZ–PT–PMN solid-solution Ferroelectric ceramics are traditionally made from powders
system,37 more-recent work in the early 1980s with PMN- formulated from individual oxides; however, the newer elec-
based relaxor ceramics has led to their successful application as trooptic materials and some of the PTC ceramics utilize chemi-
high-strain (0.1%) electrostrictive actuators38,39 and high- cal coprecipitation45–47 or hydrothermal48 techniques. The pro-
dielectric-constant (>25 000) capacitors.40 The phase diagram cessing method that one selects to prepare the powders
for this system is given in Fig. 6. The most popular specific depends, to a large extent, on cost, but even more important is
composition in this system is Pb(Mg1/3Nb2/3)O3–0.1PbTiO3, the end application. Understandably, electrooptic ceramics re-
which is PMN containing 10% PT, thus increasing the Tm (the quire higher-purity, more-homogeneous, and higher-reactivity
temperature of maximum dielectric constant for relaxors, powders than do the piezoelectric ceramics, because inhomo-
geneities can be detected optically much more easily than elec-
trically. As a result, different powder process techniques have
evolved in the two cases. Piezoelectric ceramics continue to be
prepared from the most economical, mixed-oxide (MO) pro-
cess,34 whereas the optical ceramics utilize specially developed
chemical coprecipitation (CP) processes27 involving liquid in-
organic or organometallic precursors. Although not yet fully
achieved, the trends in this area are toward the development of
one unified process that meets the objectives of both types of
materials. There is a commonality in these objectives, because
the more recent piezoelectric devices demand higher-quality
material (essentially zero porosity), and the electrooptics re-
quire a more economical process.
A flowsheet describing the essential steps for both the MO
and CP processes is given in Fig. 7. There are many steps that
are common to both methods. The essential differences be-
tween the two methods occur in the powder forming and den-
sification stages. In the MO methods, this very simply consists
of wet milling (slurry form) the individual oxides or other
compounds, such as the carbonates or nitrates that decompose
Fig. 6. Phase diagram of the PMN–PT solid-solution system.24 to the oxides during calcining (a high-temperature solid-state
806 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

are currently used with excellent success, and the choice of one
over the other is usually made on the basis of cost and conve-
nience for the end application. For details on these techniques,
it is suggested that the reader consult various excellent texts on
the subject.40,49,50
In addition to composition and powder preparation, densifi-
cation of the powder into a pore-free, fully dense ceramic el-
ement is the third area of processing that is extremely critical
to achieving a high-quality product. The flowsheet in Fig. 7
shows two methods, i.e., conventional sintering and hot press-
ing. Of these two, sintering is, by far, the oldest and most
economical method of consolidation, but it has its limitations
when it comes to achieving full density. Full density is rarely
achieved with conventional sintering of ferroelectric ceramics
unless special techniques are used to assist the sintering process
during firing. An example of this is the use of an oxygen
atmosphere for sintering lead-containing ceramics, such as
PZT and PLZT.51 With air atmosphere only, densities of ∼96%
of theoretical can be achieved, but with an oxygen atmosphere,
this value can approach 99%. Another example is the use of
excess PbO during sintering to compensate for PbO loss (vola-
tilization) as well as providing for higher densification rates via
liquid-phase sintering. When both of these techniques are used,
bulk densities approaching 100% can be achieved, as evi-
denced by the high optical transparency obtained in PLZT
9/65/35 sintered ceramics.52 Typical sintering conditions for
conventional PZT are 1250°C for 5 h with flowing oxygen and
60 h for transparent PLZT.
Although oxygen-atmosphere sintering can and does pro-
duce fully dense and transparent ceramics when the proper
procedures are used, there continues to exist a problem with
this process in consistently achieving high optical transpar-
ency. On the other hand, hot pressing is a viable method of
producing fully dense ceramics, and its worth has been proved
over several decades of experience. Slugs of PLZT as large as
Fig. 7. Flow sheet for processing of piezoelectric and electrooptic
ceramics. 150 mm (6 in.) in diameter and 25 mm (1 in.) in thickness are
regularly hot-pressed to full density and high transparency.
Typical hot-pressing conditions are 1250°C for 16 h at 14 MPa
(2000 psi).
chemical reaction) at 800°–900°C. In the CP method, the start- Other densification methods that have proved to be success-
ing materials are usually solutions that are mutually soluble in ful for ferroelectric ceramics in more recent years are (1) hot
each other, thus producing an atomically homogeneous precur- isostatic pressing, (2) vacuum sintering, and (3) a two-step
sor solution that is precipitated while blending. Because the process of sintering and then hot isostatic pressing. The two-
particle sizes of the CP powders are usually much finer than the step technique involving presintering was developed to elimi-
MO powders, (0.03–0.1 ␮m vs 1 ␮m), the CP powders are nate the need for a cladding enclosure in the final gas isostatic
more reactive and are calcined at a lower temperature, ∼500°C hot pressing step.
for 1 h. After densification, the final steps involved in the processing
Ball milling of the calcined material is necessary for both of ferroelectric ceramics (Fig. 7) are (1) slicing of the slug, (2)
types of powders to produce the required chemical and optical lapping of the slices, (3) polishing of the plates for electrooptic
homogeneity. This is a very critical step in the process, because elements, (4) electroding, and (5) evaluation of the parts for
too little milling does not produce the necessary homogeneity, further assembly to components.
and over milling increases the likelihood of contamination Some typical samples of hot-pressed and sintered PLZT and
leading to optical scattering. A common practice is to use a PZT ceramics are shown in Fig. 8 with thick (12 ␮m) films on
plastic-lined mill with high-density media (alumina or zirconia sapphire (round substrate) and glass (rectangular substrate).
balls) and a nonpolar, nonflammable milling liquid, such as The transparent part on the “sintered ceramic” label is a fully
trichloroethylene or freon TF, for the electrooptic materials; dense, oxygen-sintered, PLZT 9/65/35 ceramic.
however, distilled water is a better liquid for piezoelectrics
from a cost and environmentally preferred standpoint. Depend- (3) RAINBOW Processing
ing on the particular powder characteristics, milling times may The latest development in the processing of bulk materials
vary from 2 to 16 h. The milled powders are then thoroughly consists of the high-temperature chemical reduction of high
dried, mechanically broken up, homogenized in a V-blender, lead-containing ferroelectric wafers to produce strain-amplified
and stored for further processing. wafer actuators called RAINBOWS, an acronym for reduced
and internally biased oxide wafer.53 More specifically, this
(2) Forming and Firing (Densification) technology involves the local reduction of one surface of a
There are a variety of forming methods that have been de- ceramic wafer, thereby achieving an anisotropic, stress-biased,
veloped over the years that have been successfully used in dome or saddle-shaped configuration with significant internal
compacting the powders to a specific form or shape prior to tensile and compressive stresses that act to amplify the axial
densification. Cold pressing in a steel mold is, perhaps, the motion of the wafer and also increase the overall strength of the
oldest and most economical of these methods and, thus, is material. After reduction, the flat wafer changes its shape to
given in Fig. 7; there are, however, several more methods, one that resembles a contact lens. This is believed to be due to
including extrusion, slip casting, tape casting, roll compaction, (1) the reduction in the volume of the reduced layer (largely
screen printing, and injection molding. All of these techniques metallic lead) compared to the unreduced material, (2) the dif-
April 1999 Ferroelectric Ceramics: History and Technology 807

not; thus, one can roughly identify these two categories as

vacuum and nonvacuum techniques, respectively. The advan-
tages of the vacuum methods are (1) dry processing, (2) high
purity and cleanliness, (3) compatibility with semiconductor
integrated-circuit (IC) processing, and (4) possible epitaxial
film growth; however, these are offset by disadvantages, such
as (1) slow deposition rates, (2) difficult stoichiometry control
in ferroelectric multicomponent systems, where evaporation or
sputtering rates differ considerably, (3) high-temperature post-
deposition anneal, often required for crystallization, and (4)
high-capital equipment acquisition and maintenance costs.
The chemical techniques are usually characterized by (1)
higher deposition rates, (2) good stoichiometry control, (3)
large area, pinhole-free films, and (4) lower initial equipment
costs. These advantages, especially in the case of CVD and its
many variations, would seem to preclude the use of vacuum
methods; however, the limited availability and toxicity of some
Fig. 8. Typical examples of PLZT and PZT ceramics and films. of the ferroelectric precursors have posed some problems for
this method. Combining the advantages of excellent composi-
tional control, spin-on/spray-on/dip-coating capability, and
ferential thermal contraction between the reduced and unre- very low equipment costs, the wet chemical solution deposition
duced layers on cooling to room temperature, and (3) the vol- techniques (sol–gel and MOD) already have been quite suc-
ume change in going from the paraelectric to ferroelectric state cessful and extensively used in producing thin and thick films
at its TC. of PZT, PLZT, and many other materials. The ready availabil-
Typical steps for the RAINBOW process involve placing a ity, low cost, and water solubility of many of the precursors for
flat wafer on a graphite block, inserting the block and wafer the wet chemical methods have also significantly contributed to
into a furnace preheated to 975°C, leaving it there for 1 h and their popularity. Examples of some MOD acetate precursor
then removing it for cooling to room temperature in ∼45 min. solutions in the PZT system are shown in Fig. 9.
The net result is a monolithic (monomorph) structure consist- Some of the substrates that have been successfully used for
ing of an unreduced piezoelectric (ferroelectric) layer that is the deposition of films include silicon, platinized silicon, sap-
highly stressed, primarily in compression, and an electrically phire, magnesia, strontium titanate, silver foil, lithium niobate,
conducting reduced layer that produces the stress. These inter- gallium arsenide, fused silica, zirconia, and glass.
nal stresses have been shown to be instrumental in achieving
unusually high axial bending displacement as well as enhanced IV. Properties
load-bearing capability.54,55
(4) Thin- and Thick-Film Processing (1) Microstructure
The various techniques that are currently available for the Compositions in the PLZT system and to some degree in the
fabrication of thin films are noticeably more varied in type and PZT system, whether piezoelectric or electrooptic, are charac-
in sophistication than a couple of decades ago. Better equip- terized by a highly uniform microstructure consisting of ran-
ment and more advanced techniques have, undoubtedly, led to domly oriented grains (crystallites) tightly bonded together. An
higher-quality films and may be a primary factor in the now example of such a microstructure is given in Figs. 10(A) and
routine achievement of ferroelectricity in films as thin as 0.1 (B) for PLZT 9/65/35. The sample was polished and thermally
␮m and as thick as 22 ␮m prepared by a selection of differ- etched at 1150°C for 1 h. As is typical for most PLZT hot-
ent methods. Table II lists the major methods presently used pressed microstructures, little or no entrapped porosity is evi-
to produce ferroelectric films. Reviews of the details con- dent. This is due to the influence of the external pressure during
cerning most of these techniques are given in previous hot pressing, which aids in pore removal while the material is
publications.44,56–61 in a thermochemically active state at elevated temperatures. In
In general, there are two major categories of deposition tech- actuality, some small amount of porosity exists in all the ma-
niques for films: (1) physical vapor deposition and (2) chemical terials whether hot-pressed or atmosphere sintered; however, it
processes involving chemical vapor deposition (CVD) and is more important in the electrooptic than in the piezoelectric
chemical solution deposition. The first of these requires a materials, because porosity in the 0.5–5 ␮m diameter range is
vacuum to obtain a sufficient flux of atoms or ions capable of quite effective in scattering light. Figure 10 also shows that the
depositing onto a substrate, whereas the second usually does microstructure is very uniform, with an average grain size for
this sample of ∼8 ␮m. Piezoelectric ceramics usually possess
grain sizes in the range of 2–6 ␮m, whereas the electrooptics
cover a wider range from 2 to 10 ␮m, depending on their
Table II. Thin- and Thick-Film Deposition Techniques intended application. A uniform grain size is a highly desirable
Physical vapor deposition feature from the standpoint of performance.
Sputtering (rf magnetron, dc, ion beam)
Evaporation (e-beam, resistance, molecular beam epitaxy)
Laser ablation
Chemical vapor deposition (CVD)
MOCVD (metal–organic CVD)
PECVD (plasma-enhanced CVD)
LPCVD (low-pressure CVD)
Chemical solution deposition
Sol–gel (solution–gelation)
MOD (metalloorganic decomposition)
Melt solution deposition
LPE (liquid-phase epitaxy)
Fig. 9. Acetate precursor solutions in the PZT compositional system.
808 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Fig. 10. Typical microstructures of (A) hot-pressed electrooptic PLZT 9/65/35, (B) hot-pressed electrooptic PLZT 9/65/35 at higher magnification,
(C) hot-pressed PLZT 12/40/60 in transmitted light, and (D) chemically etched PLZT 7/65/35 in reflected light.

Domain patterns, i.e., regions of uniform and homogeneous application in dielectric, piezoelectric, pyroelectric, or elec-
spontaneous polarization within a grain or between several trooptic devices. Ferroelectrics are, in general, characterized by
grains, also can be revealed in the microstructure of a ferro- (1) higher dielectric constants (200–10 000) than ordinary in-
electric memory ceramic when transmitted light or reflective sulating substances (5–100), making them useful as capacitor
light and chemical etching techniques are used. Examples of and energy-storage materials, (2) relatively low dielectric loss
these domain patterns are shown in Figs. 10(C) and (D), re- (0.1%–7%), (3) high specific electrical resistivity (>1013 ⍀䡠
spectively, for materials with an average grain size of 8 ␮m. cm), (4) moderate dielectric breakdown (100–120 kV/cm for
Figure 10(C) shows that a predominance of 90° domains is bulk and 500–800 kV/cm for thin films), and (5) nonlinear
evident in the tetragonal PLZT 12/40/60 ceramic, whereas the electrical, electromechanical, and electrooptic behavior. Not all
domains of rhombohedral PLZT 7/65/35 in Fig. 10(D) are of these properties are optimized and realized in a given ma-
mostly 180°. The domains in 7/65/35 show up as a bilevel terial of chemical composition, and, hence, a variety of ceramic
structure, because one end of the electric dipole chemically materials are manufactured and are available from several dif-
etches faster than the opposite end. Distinctive features here are ferent companies throughout the world. A summary of typical
(1) absence of etched grain boundaries because of the fully properties for selected compositions is given in Table III.
dense structure and (2) bridging of grain boundaries by do- Small-signal (1 kHz) relative dielectric constant values for
mains, indicating little disorder at the boundaries, (these pre- several selected compositions are given in Table III. They
dominantly 180° domains are ∼3 ␮m × 15 ␮m in size). No range from a low of 225 for lead niobate to a high of 24 000 for
domains are observed in the microstructure of the 9/65/35 ma- PMN–PT (90/10). Values for the PZT and PLZT compositions
terial in Fig. 10(A), because this polished section was thermally are intermediate, ranging from 1300 for PZT-4 (a hard, A-site-
etched and not chemically etched; i.e., it was etched at 1150°C, substituted piezo material) to 5700 for a phase-boundary, re-
where domains do not exist. laxor PLZT material. The loss tangents (tan ␦) vary in value
from 0.4% to 6% for the various ceramics, and, in general,
(2) Electrical Properties the lower loss factors are associated with the lower dielectric
(A) Dielectric Properties: Because almost all of the use- constants.
ful properties of ferroelectric ceramics are related in some man- (B) Hysteresis Loops: The hysteresis loop (polarization
ner to their response with an electric field, the electrical be- versus electric field) is the single most important measurement
havior of these materials is important to their successful that can be made on a ferroelectric ceramic when characteriz-
April 1999 Ferroelectric Ceramics: History and Technology 809

(×10−10 m/V)
ing its electrical behavior. This loop is very similar to the
magnetic loop (magnetization versus magnetic field) one ob-

0.28

1.0
1.2
rC
tains from a ferromagnetic material; the very name “ferroelec-

0
0
tric” has been appropriated from this similarity, even though
there is no ferro, i.e., iron constituent, in ferroelectrics as a
(×10−16 m2/V2) major component.
Hysteresis loops come in all sizes and shapes, and, similar to

3.8
1.5
R

a fingerprint, identify the material in a very special way. There-

fore, one should become familiar with such a measurement.
Although early workers in the field of ferroelectrics utilized a
dynamic (60 Hz) measurement with a Sawyer–Tower circuit
(×10−16 m2/V2)

and an oscilloscope readout, more-recent work usually has

been done with a single-pulse or dc (∼0.1 Hz) Sawyer–Tower
1.7

3.6
m12

measurement using an X–Y plotter or computer readout.27

Typical hysteresis loops obtained from various ferroelectric
ceramic materials are illustrated in Fig. 11: (A) a linear tracing
(×10−16 m2/V2)

from a BaTiO3 capacitor, (B) a highly nonlinear loop from a

low-coercive-field (soft) memory ferroelectric such as is found
3.6

5.8

11.7
m11

in the rhombohedral region of the PZT phase diagram, (C) a

narrow, nonlinear loop obtained from a slim-loop ferroelectric
(SFE) quadratic relaxor that is located in the FE–PE boundary
region of the PLZT system, and (D) a double loop that typically
Compositions and Properties of Typical Ferroelectric Ceramics†

−0.008

−0.012
−0.008
−0.010
−0.009

−0.008
(m4/C2)

is obtained from a nonmemory antiferroelectric material in the

Q12

PSZT system.
The antiferroelectric materials are essentially nonpolar, non-
(m4/C2)

ferroelectric ceramics that revert to a ferroelectric state when

0.016

0.022
0.018
0.020
0.021

0.010
Q11

subjected to a sufficiently high electric field. Outwardly, they

differ from the SFE relaxor materials in that (1) the dielectric
constants usually are lower, (2) higher electric fields are usu-
(×10−12 m2/N)

ally required to induce the ferroelectric state, and (3) the onset
9.1
12.3
16.4
16.5
15.2

25.4
8.2
16.8

7.5
13.5
12.4
sE11

of the ferroelectric state and the return of the antiferroelectric

state are usually fairly abrupt, thus giving the loop an appear-
ance of two subloops that are positively and negatively biased.
These characteristics are shown in loop (D) in Fig. 11.
(×10−3 V⭈(m/N))

A considerable amount of information can be obtained from

a hysteresis loop. Figure 11 also shows that (1) the loop in (B)
11.4
26.1
24.8
23.1

43.1
29.5
22.2
g33

12
22
20
0
0

reveals that the material has memory, whereas the loop in (C)
indicates no memory, (2) high remanent polarization (PR) re-
lates to high internal polarizability, strain, electromechanical
coupling, and electrooptic activity, (3) for a given material, the
(×10−12 C/N)

switching field (Ec) is an indication of the grain size for a given

−78
−123
−171
−274
−241
0
−9
−51
−262

0
0

0
d31

material (i.e., lower Ec means larger grain size and higher Ec

means smaller grain size), (4) a high degree of loop squareness
usually indicates better homogeneity and uniformity of grain
size, (5) an off-centered loop from the zero voltage point (the
(×10−12 C/N)

loop is usually centered symmetrically around zero voltage)

190
289
374
593
563
0
85
127
710

235
400
682
0
0

0
d33

indicates some degree of internal electrical bias that may be

caused by internal space charge and/or aging, (6) the sharpness
Table III.

0.71
0.75
0.70

0.38
0.61
k33
0.5
0.7

0
0

0
0.36
0.58
0.65
0.58

0.07
0.46
0.72
0.34
0.47
0.62
0.65

0.65
0.6
kp

0
0

0
Data compiled from Refs. 19, 22, 24, 27, 34, 35, 62, and 63.
tan ␦

0.5
0.4
2.0
4.0

5.5
1.0
1.4
1.9
1.2
1.3
1.8
3.0
6.0
5.5
5.4
4.7
(%)
1700
1300
1700
3400
3640
24000
225
496
2590
980
1300
1850
3400
5700
5500
4940
5100
K
115
328
365
193
185
40
570
420
160
245
145
150
110
80
75
100
85
(°C)
TC
Density
(g/cm3)
5.7
7.5
7.8
7.5
7.6
7.6
6.0
4.5
7.8
7.8
7.7
7.8
7.8
7.8
7.8
7.8
7.8
PMN–PT (65/35)
PMN–PT (90/10)

(Na0.5K0.5)NbO3

PLZT 9.5/65/35
PLZT 7.6/70/30
PLZT 12/40/60
PLZT 7/60/40
PLZT 8/40/60
PLZT 7/65/35
PLZT 8/65/35
PLZT 9/65/35

PLZT 8/70/30
Composition

PbNb2O6
PZT-5A
PZT-5H
BaTiO3

Fig. 11. Typical hysteresis loops from various ferroelectric ceramics:

PZT-4

(A) BaTiO3 capacitor, (B) soft (easily switchable) PZT, (C) PLZT
†

8.6/65/35 relaxor, and (D) PSZT antiferroelectric material.

810 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

of the loop tips indicates a high electrical resistivity (>109 constant at the MPB, (3) a larger number of reorientable po-
⍀䡠cm), (7) high induced polarization in relaxor materials indi- larization directions existing in the MPB mixed-phase region,
cates high electrostriction strain and high electrooptic coeffi- and (4) a maximum in mechanical compliance in the boundary
cients, (8) the slope of the P–E loop at any point along the loop region, permitting maximum domain reorientation without
is equal to the large-signal dielectric constant, (9) the opening physically cracking.
up of the loop of a SFE relaxor material can indicate nonohmic Also included in Table III are some typical electrostrictive Q
contact between the electrodes and the ceramic, and (10) a and m values for representative compositions. Table III shows
sudden large change in “apparent” polarization is usually a that most Q11 coefficients are in a rather narrow range of
result of incipient dielectric breakdown. Remanent polariza- 0.010–0.022 m4/C2, as are the Q12 coefficients in a range of
tions for most of the lead-containing ferroelectrics typically 0.008–0.012 m4/C2. Also given are the Q values of two PLZT
vary from 30 to 40 ␮C/cm2, whereas the coercive fields vary ferroelectric compositions (7/65/35 and 8/65/35) for compari-
over quite a wide range, from ∼2 kV/cm to near electrical son, pointing out the observation that the Q coefficients are
breakdown (∼125 kV/cm), depending on the type of dopants similar in magnitude regardless of the ferroelectric or nonfer-
and modifiers added. roelectric nature of the material. This is because the Q coeffi-
The strains associated with two of these materials (i.e., fer- cient relates the resulting strain to the electrically induced po-
roelectric and SFE) on traversing their hysteresis loops are larization, regardless of whether the material has permanent
given in Fig. 12. In the ferroelectric case, the switching strain polarization.
accompanying the polarization reversal process results in the The m coefficients, on the other hand, relate the strain to the
familiar “butterfly” loop, with the remanent strain state in the electric field; hence, their values vary more widely, ranging
center of the loop (point O). Positive voltage then results in a from 1.7 × 10−16 to 11.7 × 10−16 m2/V2.
longitudinal expansion of the ceramic, whereas a negative volt- (D) Pyroelectric Properties: Although the pyroelectric
age (less than the coercive field) results in a longitudinal con- effect in crystalline materials has been known for many cen-
traction. This is known as the linear strain effect in piezoelec- turies, it has been within only the last four decades that this
tric materials and does not involve domain switching. For the effect has been studied in ferroelectric ceramics.66–69 As men-
SFE relaxor case, there is no remanent strain when the electric tioned previously, this effect occurs in polar materials and is
field is not applied, because, in this case, the rest position of the manifested in a change in polarization as a function of tem-
ion is in the center of the unit cell. However, when the field is perature. This results in a reduction of the bound charge re-
applied, ionic movement (polarization) and strain occur simul- quired for compensation of the reduced dipole moment on
taneously, both being dependent upon the strength of the field. increasing temperature and vice versa on decreasing tempera-
Because the sign of the strain produced (positive for elonga- ture; thus, the change in voltage on the material’s electrodes is
tion) is the same regardless of the polarity of the field, this is a measure of the change in the material’s polarization due to
the electrostrictive effect mentioned previously. absorbed thermal energy. A common figure-of-merit for pyro-
(C) Piezoelectric and Electrostrictive Properties: Com- electrics is
positions within the PZT and PLZT systems possess some of
the highest electromechanical coupling coefficients attainable p
FOM = (13)
in ceramic materials. Some typical values of kp, k33, d33, d31, c共K tan ␦兲1 Ⲑ 2
and g33 for these materials are given in Table III with BaTiO3
and the niobates. Maximum values of kp (0.72) and d33 (710 × where p is the pyroelectric charge coefficient, c the specific
10−12 C/N) are found in the soft (easily switchable) PLZT heat, and tan ␦ the dielectric loss tangent. Maximizing the
composition 7/60/40. This composition is located within the performance of a material then involves selecting a ceramic
morphotropic phase-boundary region separating the ferroelec- with a high pyroelectric coefficient and low specific heat, di-
tric rhombohedral and tetragonal phases. Over the years, there electric constant, and dielectric loss factor. This is difficult to
has been considerable speculation concerning the reasons for achieve in a given material, and, most often, its performance is
this maximum in coupling at the MPB.64,65 These may be limited by the dielectric loss, which is reflected in a poor sig-
summarized as being due to (1) the existence of a mixture of nal-to-noise ratio.
phases at the boundary, (2) a concurrent maximum in dielectric Two families of ceramics have dominated this area of en-
deavor: PZT and BST (barium strontium titanate) materials.
However PLZT and PMN are also considered viable candi-
dates. The former two materials are considered ferroelectric
thermal detectors (absorbed energy generating the temperature-
dependent change in polarization), whereas the latter two, as
well as BST, can be considered dielectric bolometers (electri-
cally induced, temperature-dependent change in dielectric con-
stant materials).68 Ceramics, in many cases, are considered
better choices for thermal imaging applications than crystalline
materials with higher pyroelectric coefficients because of their
lower cost, availability, ease of processing, and good stability.
These materials in bulk and thin-film forms are used in com-
mercial products for law-enforcement, night surveillance, and
security applications.
(E) Optical and Electrooptic Properties: Unlike the PZT
ceramics and other ferroelectric materials that are opaque, the
most outstanding feature of the PLZT materials is their high
optical translucency and transparency. Optical transparency is
both a function of the concentration of lanthanum and the Zr/Ti
ratio with a maximum in transparency occurring along the
FE–PE phase boundary and beyond, until mixed phases pro-
duce opacity (see Fig. 5). For example, the 65/35 Zr/Ti ratio
compositions are most transparent in the lanthanum range from
8 to 16 at.%, whereas the 10/90 compositions are similarly
Fig. 12. Hysteresis loops and longitudinal strain curves for (A) fer- transparent in the 22% to 28% range.
roelectric memory ceramic and (B) SFE nonmemory relaxor ceramic. A typical transmission curve for a 9/65/35 composition is
April 1999 Ferroelectric Ceramics: History and Technology 811

given in Fig. 13. The material is highly absorbing below 0.37 V. Applications
␮m, which is the commonly accepted value for the onset of The applications for ferroelectric ceramics are manifold and
high absorption in the bulk material. For thin films, this value pervasive, covering all areas of our workplaces, homes, and
is closer to 0.35 ␮m. A fairly constant optical transmission of automobiles. Similar to most materials, the successful applica-
∼65% occurs throughout the visible spectrum from 0.5 ␮m to tion of these piezoelectric, pyroelectric, ferroelectric, elec-
the near infrared at 6.5 ␮m (see inset). Beyond this, absorption trostrictive, and electrooptic ceramics and films are highly de-
again begins to take place, and, at 12 ␮m, the material is, once pendent on the relative ease with which they can be adapted to
again, fully absorbing. The high-surface-reflection losses useful and reliable devices. This is, to a great extent, the reason
(∼31% for two surfaces) shown in Fig. 13 are a function of the that they have been so successful over the years in finding an
high index of refraction (n ⳱ 2.5) of the PLZT. increasing number of applications. Their simplicity, compact
Four common types of electrooptic effects have been found size, low cost, and high reliability are very attractive features to
to be operative in ferroelectric materials in general and in the design engineer. Many general category applications for
PLZT ceramics in particular: (1) quadratic, Kerr, and birefrin- bulk and film electroceramics are given in Fig. 14. As indicated
gent effects, (2) depolarization nonmemory scattering, (3) lin- in Fig. 14, some of these applications are more appropriate for
ear, Pockels, and birefringent effects, and (4) memory scatter- bulk materials, some for films, and some for both bulk and
ing. The first two types utilize relaxor-type, 9/65/35 materials films. Although there always will be a demand for bulk de-
with linearly polarized light; the third type uses a high coercive vices, it is certainly obvious that the trend in the industry is
field, tetragonal, memory material, such as 12/40/60, with po- toward film devices. The reasons for this include (1) lower
larized light; and the fourth type commonly uses a low coercive operating voltages, (2) size and weight compatibility with in-
field, rhombohedral, memory material, such as 7/65/35, and tegration trends, (3) better processing compatibility with sili-
does not use polarizers, but, rather, relies on the variable-angle con technology, (4) ease of fabrication, and (5) lower costs
scattering of light from different polarized areas to achieve a through integration.
spatially varying image in the ceramic. Contrast ratios as high
as 3000/1 can be attained with polarized light, whereas these (1) Capacitors
ratios are limited to <50/1 for schemes involving nonpolarized, One category of applications for ferroelectric-type materials
scattered light. Specific properties of the more-common PLZT is that of high-dielectric-constant capacitors, particularly
electrooptic compositions are listed in Table III. MLCs. MLCs are extremely important to our everyday lives in
PLZT materials are also known to possess many special that they are essential to all of our currently produced elec-
photosensitive phenomena that are directly linked to their mi- tronic components, and, as such, they constitute a significant
crostructural, chemical, electronic, and optical properties, in- portion of the multibillion dollar electronic ceramics business
cluding (1) photoconductivity, (2) photovoltaic properties, (3) as a whole. Most ceramic capacitors are, in reality, high-
photo-assisted domain switching, (4) ion-implantation- dielectric-constant ferroelectric compositions which have their
enhanced photosensitivity, (5) photochromic effects, (6) pho- ferroelectric (hysteresis loop) properties suppressed with suit-
tomechanical (photostrictive) behavior, (7) photorefractive ef- able chemical dopants while retaining a high dielectric constant
fects, and (8) photoexcited space charge phenomena.70,71 over a broad temperature range. BaTiO3 was historically the
Although materials with such a multitude of properties and first composition used for high-dielectric-constant capacitors,
special effects hold promise for many new applications for the and it (or its variants) remains the industry standard; however,
future, it should also be remembered that these same effects lead-containing relaxors such as PMN and PZN are making
can, and often do, limit their application. inroads.72 In tune with ever-shrinking electronic components in

Fig. 13. Optical transmission characteristics of electrooptic PLZT (9/65/35).

812 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Fig. 14. Applications of bulk and film ceramic electronic materials.

this age of integration, capacitor techniques have trended to- their strain (displacement) is limited to ∼0.5% or less. On the
ward (1) more-sophisticated tape-casting procedures, (2) sur- other hand, maximum displacement of several tens of percent
face-mount MLCs, and (3) fired layer thicknesses approaching can be achieved with displacement-amplifying means, such as
4 ␮m. MLCs, 0.5 mm × 1mm and several hundred layers thick, composite (flextensional structure, Moonie) or bender (uni-
are now produced with capacitances of several microfarads. morph, bimorph, RAINBOW) structures, but this is usually
Tape-casting methods are now reaching their practical limit, accomplished at the expense of considerably less force genera-
and thin-film deposition techniques are being explored. Typical tion, greater complexity, and higher cost. In most cases, the
applications include general-use discrete and MLCs, voltage- actuators are operated with electric fields <10 kV/cm for lon-
variable capacitors, and energy-storage capacitors.28,40 gevity and reliability. However, even such modest fields can
By far the largest majority of applications for electro-active result in rather high voltages (>1000 V) if the actuator is rela-
materials occurs in the area of piezoelectric ceramics. In this tively thick (V ⳱ Et); thus, the multilayer technology devel-
category, the ceramics are usually poled once at the factory, oped for capacitors is often used to reduce the operating volt-
and no polarization reorientation takes place after that through- age below 100 V.
out the life of the device. These devices can be divided into Although unimorph and bimorph structures have been suc-
four different groups, as given in Fig. 15. Two of these groups
are as mentioned previously, i.e., motors and generators; how-
ever, the third category involves the use of combined motor
and generator functions in one device, and the fourth category
includes devices operated at higher frequencies, i.e., at reso-
nance. Because of the more-recent interest in electrically bi-
ased electrostrictive devices that act as electrically tunable pi-
ezoelectric components, some of the specific applications in
Fig. 15 are also now being developed with electrostrictive ma-
terials. Examples of ceramics that are utilized in a variety of
piezoelectric and electrostrictive applications, both large and
small, are shown in Fig. 16.
Figure 15 also shows that the number of applications for
piezoelectrics as motors is quite numerous. This is particularly
true for the whole family of micro- and macro-piezoelectric
actuators. The micro-devices are considered to be those that
utilize the basic piezoelectric strain of the ceramic (measured in
micrometers), whereas the macro-devices are those that use a
displacement-amplifying mechanism to enhance the fundamen-
tal strain (measured in millimeters). This is explained more
thoroughly in Table IV, which lists all of the current ceramic
actuator technologies and includes some of their important
characteristics. Table IV shows that a variety of direct exten-
sional configurations, composite flextensional structures, and
bending-mode devices are used to achieve a mechanical output.
Maximum stress generation (40 MPa) or loading capability is
noted for all of the direct extensional devices, including the Fig. 15. Piezoelectric and electrostrictive applications for ferroelec-
piezoelectrics, electrostrictors, and antiferroelectrics; however, tric ceramics.
April 1999 Ferroelectric Ceramics: History and Technology 813

Fig. 16. Variety of ferroelectric ceramics used in piezoelectric and electrostrictive applications, such as sonar, accelerometers, actuators, and
sensors. (Photograph courtesy of EDO Western.)

cessfully applied to many devices over the past four decades, mode of operation while sustaining loads of 1 kg. Maximum
their inability to extend the force–displacement envelope of displacements of >1 mm can be achieved with wafers (32 mm
performance has led to a search for new actuator technologies. diameter), thinner than 0.25 mm when operating in a saddle
One such device developed in the early 1990s is the Moonie— mode. Prototypes of RAINBOW pumps, speakers, optical de-
so named because of its crescent-shaped, shallow cavities on flectors, vibratory feeders, relays, hydrophones, switches, plat-
the interior surfaces of the end caps (see Table IV), which are form levelers, sensor and actuator arrays, and toys have been
bonded to a conventionally electroded piezoelectric ceramic demonstrated; however, no commercial products have been yet
disk. When the ceramic is activated electrically, the shallow been produced.53,76 Some examples of these different types of
cavities permit the end caps to flex, thus converting and am- piezoelectric devices are included in Fig. 17.
plifying the radial displacement of the ceramic into a large A novel type of bimorph application of somewhat recent
axial motion at the center of the end caps. Advantages of the vintage is the optomechanical (photostrictive) actuator. The
Moonie include (1) a factor of 10 enhancement of the longi- photostrictive behavior is a result of a combined photovoltaic
tudinal displacement, (2) an unusually large d33 coefficient effect (wherein light produces a voltage in the ceramic) and a
exceeding 2500 pC/N, and (3) enhanced hydrostatic re- piezoelectric effect (wherein this voltage produces a strain in
sponse.73,74 Recent improvements in the basic Moonie design the material via the converse piezoelectric effect). PLZT ce-
have resulted in an element called a Cymbal, a device that ramics with donor-type doping exhibit large photostrictive ef-
possesses more-flexible end caps, resulting in higher displace- fects when irradiated with high-energy, near-ultraviolet light. A
ment.75 Applications include transducer arrays, medical imag- bimorph configuration with no connecting wires has been used
ing transducers, and hydrophones. to demonstrate prototypes of a photo-driven relay and a remote
Another device recently developed to increase the force– micro-walking device, and a photophone of the future has been
displacement performance of a piezoelectric actuator is the envisioned.77,78
RAINBOW. In its most basic sense, a RAINBOW can be
thought of as a prestressed, axial-mode bender similar in op- (2) Explosive-to-Electrical Transducers (EETs)
eration to the more conventional unimorph bender. Unlike the Studies on the stress-induced depoling of ferroelectric ce-
unimorph and Moonie, which are composite structures, the ramics were initiated in the mid-1950s, which culminated in
RAINBOW is a monolithic monomorph that is produced from the development of one-shot power supplies that made use of
a conventional, high-lead-containing piezoelectric ceramic by this effect. This depoling behavior is optimum (i.e., maximum
means of a high-temperature, chemical reduction reaction. As output in the shortest period of time) for ferroelectric compo-
mentioned previously, this process produces significant inter- sitions located along the boundary between the polar ferroelec-
nal compressive and tensile stresses that are instrumental in tric phase and the nonpolar antiferroelectric phase, such as
achieving its unusually high displacement characteristics. Dis- shown in the gray area of the phase diagram of Fig. 5. Although
placements as high as 0.25 mm for a 32 mm diameter × 0.5 mm this depoling does occur somewhat more slowly under hydro-
thick wafer have been achieved for these devices in a dome static pressure, when it is accomplished in an extremely fast
814 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Table IV. Current Ceramic-Actuator Technologies

Maximum Actuator Actuator
stress generated movement Actuator displacement
†
Type Configuration (MPa) (with voltage applied) type‡ (%)§

Monolithic (d33 mode) 40 Expansion P 0.40

Monolithic (d31 mode) 40 Contraction P −0.15

Monolithic (s11 mode) 40 Expansion E 0.28

Monolithic (s12 mode) 40 Contraction E −0.09

Monolithic (s11 mode) 40 Expansion A 0.50

Monolithic (s12 mode) 40 Expansion A 0.08

Composite structure
(d33 mode) (flextensional) 10 Contraction P/E −1.0

Composite structure 0.028 Expansion P/E 1.3

(d33/d31) (Moonie)

Unimorph (bender) 0.006 Expansion/contraction P/E 10

Bimorph (bender) 0.006 Expansion/contraction P/E 20

Rainbow monomorph (bender) 35 (D)

0.02 Expansion/contraction P/E/A
450 (S)
†
V is voltage and D is actuator displacement. ‡P is piezoelectric, E is electrostrictor, and A is antiferroelectric. §D is dome mode and S is saddle mode; displacement at ±10 kV/cm
based on thickness.

mode via explosive shock waves or projective impact, useful ing pattern in which both phases are three-dimensionally self-
electrical pulses of a few hundred kilovolts or kiloamps lasting connected). Some of these connectivity patterns are particu-
for many microseconds can be obtained. These one-shot power larly well suited for decoupling the longitudinal and transverse
supplies have found many uses in military applications.79,80 piezoelectric effects, such that materials with significantly en-
hanced (up to a factor of 100 or more) piezoelectric properties
(3) Composites are possible. Moreover, a ceramic–polymer composite offers
Piezoelectric composites represent one of the latest technolo- distinct advantages, such as a wide range of acoustic imped-
gies developed for engineering the last bit of high performance ance matching, broad bandwidth, low electrical losses, and, for
from a piezoelectric transducer. When one deliberately intro- medical ultrasound applications, send–receive capability in a
duces a second phase in a material, connectivity of the phases compact package. Considerable engineering ingenuity has been
is a critical parameter. There are 10 connectivity patterns pos- demonstrated in designing, fabricating, and packaging the
sible in a two-phase solid, ranging from 0–0 (unconnected many types of diphasic structures. Major applications include
three-dimensional checkerboard pattern) to 3–3 (interpenetrat- hydrophones, sensors, and medical ultrasonics.81–84
April 1999 Ferroelectric Ceramics: History and Technology 815

mented transmissive and reflective displays.27,34 Several com-

mercial and military applications for PLZT shutters and linear
gate arrays are given in Fig. 18.
(5) Films
By far the largest number of applications in ferroelectric
ceramics remains associated with bulk materials, but a trend
toward thin and thick films for some of these applications has
recently developed and is steadily increasing in intensity. Aside
from the obvious advantages, such as smaller size, less weight,
and easier integration to IC technology, ferroelectric films offer
additional benefits, including (1) lower operating voltage, (2)
higher speed, and (3) the ability to fabricate unique micro-level
structures. Equally important, but not as obvious, is the fact
that many materials that are difficult, if not impossible, to
fabricate to a dense ceramic as a bulk material are relatively
easy to produce as films. Moreover, the sintering temperatures
Fig. 17. Examples of PZT, PLZT, and PMN piezoelectric and elec- of the films are usually hundreds of degrees celsius lower than
trostrictive devices: (starting at upper right and going clockwise) Mo- that of the bulk, and this often can be the deciding factor in a
torola tweeter, Triangle gas-grill lighter, Motorola bimorph, Murata successful design and application.57,59–61
intermediate frequency resonators, Morgan Matroc ultrasonic cleaner
ceramics, Aura RAINBOW ceramics, Itek PMN actuator, ferroelectric
The several most important film applications are included in
film memory, Kodak PLZT E/O device, RAINBOW mouse toy ac- Fig. 14. Here again, some applications are exclusively desig-
tuator, Moonie actuators, Radio Shack buzzer, and unimorphs. nated for films while others are mutually shared with bulk
materials. Applications for thick films (2–20 ␮m) include elec-
trooptic and some piezoelectric devices, whereas the remaining
applications (i.e., capacitors, infrared sensors, memories, buffer
(4) Electrooptics layers, integrated optics, and antireflection coatings) involve
Since the late 1960s, when transparent PLZT materials were thin films ranging in thickness from 0.2 to 2 ␮m.85 The reason
first developed, ferroelectric ceramics have been thoroughly for the thicker films in the former cases is to obtain maximum
researched, and their characteristics studied to the point that strain or electrooptic output from the films without having to
they have now taken their place alongside single crystals as resort to electric fields too near to dielectric breakdown.86 This
legitimate candidates for electrooptic applications. Compared is not a problem in the latter applications, because the films are
to single crystals, ferroelectric ceramics are generally, but not subjected only to low voltages (<10 V) or to no voltage in their
always, less transparent, less uniform on a microscopic scale operational environment.
because of their polycrystalline nature, and somewhat less op- Large-scale manufacturing is presently underway to incor-
timum in their electrooptic and hysteretic properties. On the porate ferroelectric films as storage and bypass capacitors in
other hand, electrooptic ceramics have several specific charac- and other IC circuitry. In a conventional DRAM (dynamic
teristics that make them well suited for a variety of electrooptic random-access memory) computer memory application, one
applications. Among these characteristics are (1) small areas SiO2 capacitor is in series with every switching transistor; how-
(on the order of the grain size, from 1 to 10 ␮m) that can be ever, as the memories become more dense, the area taken up by
electrically switched independently of other adjacent areas, (2) the low-dielectric-constant SiO2 capacitor is too large, and new
switched areas that are stable with time in memory materials or materials with higher dielectric constants must be substituted.
stable with applied electric field in nonmemory materials, (3) A ferroelectric DRAM (FEDRAM) film, because of its much
light transmission characteristics of the switched areas that higher dielectric constant (∼1000 vs 4), occupies much less
depend on the thoroughness and direction of switching or pol- wafer area than the normal SiO2 capacitor, thus allowing much
ing, (4) switched areas that exhibit either electrically variable greater capacity memories to be fabricated on a given silicon
light scattering behavior or electrically variable birefringence, wafer.59 As memories become more dense in the future, the
and (5) ceramics that can be hot-pressed or sintered in a variety transition to ferroelectric films will become a necessity, and
of sizes and shapes with a high degree of optical uniformity on operating voltages for these memories will continue to decrease
a macroscopic scale. toward 1 V. At present, BST film capacitors are the top con-
Thin polished plates of PLZT, such as 9/65/35, when used in tenders for these applications.
conjunction with linearly polarized light, make excellent wide- The development of ferroelectric random access memory
aperture electrooptic shutters, linear gate arrays, and color fil- (FERRAM) films for nonvolatile memory applications, such as
ters. Their fast response (in the low microsecond range), thin computer random-access memories, smart cards, and radio-
profile, wide viewing angle, and wide temperature range of frequency identification tags, is presently underway and has
operation (−40°C to +80°C) are highly desirable characteristics reached modest production levels for specific niche applica-
for most devices; however, these are offset by the disadvan- tions.59 The issues faced in nonvolatile memories are much
tages of (1) low ON-state transmission of ∼15% as a result of more challenging than those involving simple capacitors, be-
the polarizers and (2) high operating voltages (∼350 V) that are cause, in the nonvolatile case, the two memory states consist of
required to reach a full ON condition. Despite the low ON-state the two polarization states (polarization up and polarization
transmission characteristics, high contrast ratios of 3000/1 are down in Fig. 3) in the film, one being a 0 and the other being
easily achieved when high-efficiency polarizers are used. a 1. In this case, the ferroelectric polarization is constantly
The polarizer–PLZT configuration usually used in a shutter being switched, and this immediately raises the issue of switch-
device is shown in Fig. 4. Vacuum-deposited or grooved, in- ing fatigue. FERRAM films of several compositions, including
terdigital electrodes (not shown) are commonly used to de- PZT, PLZT, and SrBi2Ta2O7 (SBT is a layer-type ferroelectric,
crease the operating voltage by reducing the gap between the see Table II), are actively being investigated. Among these,
positive and negative finger electrodes. These finger electrodes SBT is, perhaps, the material of choice in regard to fatigue.
are thin enough (<0.075 mm) so that they can be optically SBT materials exhibiting little (<10%) fatigue out to 1012
defocused and the imaging preserved. Such an arrangement is switching cycles have been developed; however, this value
used in the eye-protective systems developed by the military, in must be increased to ∼1014 cycles before these films can be
linear light gate arrays for printers and processors, and in seg- used in large-scale applications. Electrode–ceramic interac-
816 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4

Fig. 18. Commercial and military applications of PLZT electrooptic ceramics: (starting at upper right and going clockwise) EEU-2P flyers goggles
(Photograph courtesy of Sandia National Laboratories.), B1-B cockpit viewing port (Photograph courtesy of Sandia National Laboratories.), film
writers (Photograph courtesy of LVT.), offset image setter (Photograph courtesy of Steiger.), and premier image enhancement system (Photograph
courtesy of Eastman Kodak.).

tions, resistance degradation, and memory imprint (persistent composites and displacement-amplifying techniques and mate-
memory, i.e., the resistance to switch out of a given memory rials will proliferate in a continuing effort to widen the force–
state) also have been identified as major problems in these displacement envelope of performance. These devices, too,
materials, and satisfactory solutions to these issues are critical will become smarter and smarter as the applications demand.
to their success. Brought on by the need for higher-capacity memories, ex-
panded data processing capability, and smarter devices, the
VI. Future Prospects direction set a few years ago for ferroelectric films is expected
to continue and broaden in scope. Thin- and thick-film tech-
Present market trends continue to show that the future for nologies alike will also share in the current trend toward com-
ferroelectric ceramics is bright and continues to get even posite and graded structures with specifically engineered, and
brighter as the transition is made from passive to electrically often unique, properties. Multiply deposited layers of different
active “smart” and “very smart” materials. In this regard, a materials, or graded layers of the same material, are now
smart material senses a change in the environment and, using achievable with most conventional film deposition processes
an external feedback control system, makes a useful response, on a micro scale, and this will be more commonplace in the
as in a combined sensor/actuator ceramic. A very smart mate- future on a nano scale.88–91
rial senses a change in the environment and responds by react- Because thin- and thick-film technologies generally do not
ing and tuning (self-controlling) one or more of its properties to limit, but rather enhance, the portfolio of materials to be used
optimize its behavior. An example of the smart type is a pi- in various applications, it is expected that a variety of materials
ezoelectric ceramic and of the very smart type is a nonlinear, will continue to be studied, but there will be a narrowing down
electrostrictive relaxor. Multifunctionality is a key concept of to fewer serious candidates of known behavior in order to bring
these materials that will be exploited with all the ingenuity that the devices in development to the marketplace. Undoubtedly,
design engineers can muster.87 BST, PZT, PLZT, PMN, and SBT are destined to be leading
In the future, more and more applications for nonlinear, candidates in this arena. Regarding film deposition techniques,
electrostrictive relaxor materials, such as PMN and PLZT, will at this stage in the development of the films, it is very difficult
emerge as the relentless drive toward miniaturization and in- to judge which film deposition technique will emerge as the
tegration continues. Indeed, this very trend will also encourage favorite; however, because several methods have been used
more materials research efforts to develop better ferroelectric successfully, it is most likely that several methods will survive,
and electrostrictive ceramics. and a specific method selected will be dictated by cost and the
As niche applications become more prevalent in the future, application.
April 1999 Ferroelectric Ceramics: History and Technology 817

Because of their intrinsic dielectric nature and large number Ceramic Materials for Electronics. Edited by R. C. Buchanan. Marcel Dekker,
of interactive and electrically variable properties, ferroelectric New York, 1986.
35
G. H. Haertling and C. E. Land, “Hot-Pressed (Pb,La)(Zr,Ti)O3 Ferroelec-
ceramics are destined to figure prominently in the future. Bulk, tric Ceramics for Electrooptic Applications,” J. Am. Ceram. Soc., 54 [1] 1–11
thick-film, and thin-film forms of these materials have now (1971).
36
proved their worth, and, together, they will constitute a strong G. A. Smolenskii and A. I. Agranovskaya, “Dielectric Polarization of a
portfolio of materials for future applications in electronics. Number of Complex Compounds,” Sov. Phys.–Solid State (Eng. Transl.), 1 [10]
1429–37 (1960).
37
H. Ouchi, K. Nagano, and S. Hayakawa, “Piezoelectric Properties of
Pb(Mg1/3Nb2/3)O3–PbTiO3–PbZrO3 Solid-Solution Ceramics,” J. Am. Ceram.
Soc., 48 [12] 630–35 (1965).
References 38
L. E. Cross, S. J. Jang, and R. E. Newnham, “Large Electrostrictive Effects
1
J. Valasek, “Piezoelectric and Allied Phenomena in Rochelle Salt,” Phys. in Relaxor Ferroelectrics,” Ferroelectrics, 23, 187–92 (1980).
Rev., 17, 475–81 (1921). 39
S. Nomura and K. Uchino, “Recent Applications of PMN-Based Electric-
2
G. Busch, “Early History of Ferroelectricity,” Ferroelectrics, 74, 267–84 (1987). tors,” Ferroelectrics, 50, 197–202 (1983).
3
W. Kanzig, “History of Ferroelectricity, 1938–1955,” Ferroelectrics, 74, 40
M. Kahn, D. Burks, I. Burn, and W. Schulze, “Ceramic Capacitor Tech-
285–91 (1987). nology”; pp. 191–274 in Electronic Ceramics. Edited by L. M. Levinson. Mar-
4
L. E. Cross and R. E. Newnham, “History of Ferroelectrics”; pp. 289–305 in cel Dekker, New York, 1988.
Ceramics and Civilization, Vol. III, High-Technology Ceramics—Past, Present, 41
M. F. Yan, H. C. Ling, and W. W. Rhodes, “Preparation and Properties of
and Future. American Ceramic Society, Westerville, OH, 1987. PbO–MgO–Nb2O5 Ceramics Near the Pb(Mg1/3Nb2/3)O3 Composition,” J.
5
G. Busch, “How I Discovered the Ferroelectric Properties of KH2PO4,” Mater. Res., 4 [4] 930–44 (1989).
Ferroelectrics, 71, 43–47 (1987). 42
S. L. Swartz and T. R. Shrout, “Fabrication of Perovskite Lead Magnesium
6
J. Fousek, “Ferroelectricity: Remarks on Historical Aspects and Present Niobate,” Mater. Res. Bull., 17, 1245–50 (1982).
Trends,” Ferroelectrics, 113, 3–20 (1991). 43
M. Sayer and K. Sreenivas, “Ceramic Thin Films: Fabrication and Appli-
7
H. Thurnauer, “Reflections,” Am. Ceram. Soc. Bull., 56 [10] 861–66 (1977). cations,” Science (Washington, DC), 247, 1056–60 (1990).
8
W. W. Coffeen, “Ceramic Capacitors—Past, Present, and Future,” Ceram. 44
G. H. Haertling, “Ferroelectric Thin Films for Electronic Applications,” J.
Ind.; published in three parts: Nov. 1974, Feb. 1975, and June 1975. Vac. Sci. Technol. A, 9 [3] 414–20 (1991).
9
B. M. Wul and I. M. Goldman, “Dielectric Constants of Titanates of Metals 45
O. Saburi, “Properties of Semiconductive Barium Titanates,” J. Phys. Soc.
of the Second Group,” Dokl. Akad. Nauk SSSR, 46, 154–57 (1945); Compt. Jpn., 14 [9] 1159–74 (1959).
Rend. Acad. Sci. URSS, 49, 139–42 (1945). 46
G. H. Haertling and C. E. Land, “Recent Improvements in the Optical and
10
A. von Hippel, R. G. Breckenridge, F. G. Chesley, and L. Tisza, “High Electrooptic Properties of PLZT Ceramics,” Ferroelectrics, 3, 269–80 (1972).
Dielectric Constant Ceramics,” Ind. Eng. Chem., 38 [11] 1097–109 (1946). 47
K. S. Mazdiyasni, “Fine Particle Perovskite Processing,” Am. Ceram. Soc.
11
R. B. Gray, “Transducer and Method of Making Same,” U.S. Pat. No. Bull., 63 [4] 591–94 (1984).
2 486 560, 1949. 48
W. J. Dawson, “Hydrothermal Synthesis of Advanced Ceramic Powders,”
12
H. Jaffe, “Piezoelectric Ceramics,” J. Am. Ceram. Soc., 41 [11] Part II, Am. Ceram. Soc. Bull., 67 [10] 1673–38 (1988).
494–98 (1958). 49
W. D. Kingery (Ed)., Ceramic Fabrication Processes; pp. 1–235. MIT
13
B. Jaffe, W. R. Cook Jr., and H. Jaffe; pp. 1–5 in Piezoelectric Ceramics; Technology Press and Wiley, New York, 1960.
Ch. 1, pp. 1–5. Academic Press, New York, 1971. 50
W. S. Young, “Multilayer Ceramic Technology”; pp. 403–24 in Ceramic
14
A. von Hippel, “Ferroelectricity, Domain Structure, and Phase Transitions Materials for Electronics. Edited by L. M. Levinson. Marcel Dekker, New
of Barium Titanate,” Rev. Mod. Phys., 22 [3] 221–37 (1950). York, 1988.
15
K. Hardtl, “Ferroelectrics”; in Kirk-Othmer: Encyclopedia of Chemical 51
T. F. Murray and R. H. Dungan, “Oxygen Firing Can Replace Hot Pressing
Technology, Vol. 10, 3rd ed.; pp. 1–30. Wiley, New York, 1980. for PZT,” Ceram. Ind., 82 [6] 74–77 (1964).
16
W. G. Cady, Piezoelectricity; pp. 1–20. McGraw-Hill, New York, 1946. 52
G. S. Snow, “Improvements in Atmosphere Sintering of Transparent PLZT
17
“IEEE Standard Definitions of Primary Ferroelectric Terms,” ANSI/IEEE Ceramics,” J. Am. Ceram. Soc., 56 [9] 479–80 (1973).
Std. 180-1986, pp. 1–21, IEEE, New York, 1986. 53
G. H. Haertling, “Rainbow Ceramics—A New Type of Ultra-High Dis-
18
“IRE Standards on Piezoelectric Crystals: Measurements of Piezoelectric placement Actuator,” Am. Ceram. Soc. Bull., 73 [1] 93–96 (1994).
Ceramics, 1961,” Proc. IRE, 49 [7] 1161–69 (1961). 54
C. Elissalde, L. E. Cross, and C. A. Randall, “Structure–Property Relations
19
H. Jaffe and D. A. Berlincourt, “Piezoelectric Transducer Materials,” Proc. in a Reduced and Internally Biased Oxide Wafer (RAINBOW) Actuator Ma-
IEEE, 53 [10] 1372–86 (1965). terial,” J. Am. Ceram. Soc., 79 [8] 2041–48 (1996).
20
A. Ballato, “Piezoelectricity: Old Effect, New Thrusts,” IEEE Trans. Ul- 55
G. Li, “Influence of Internal Stress on the Electromechanical Properties of
trason., Ferroelectr., Freq. Control, 42 [5] 916–25 (1995). PLZT Stress-Biased (RAINBOW) Ceramic Actuators”; Ph.D. Dissertation,
21
S. J. Jang, K. Uchino, S. Nomura, and L. E. Cross, “Electrostrictive Be- Clemson University, Clemson, SC, 1995.
havior of Lead Magnesium Niobate-Based Ceramic Dielectrics,” Ferroelectrics, 56
K. D. Budd, S. K. Dey, and D. A. Payne, “Sol–Gel Processing of PbTiO3,
27, 31–34 (1980). PbZrO3, PZT, and PLZT Thin Films,” Br. Ceram. Proc., 36, 107–21 (1985).
22
M. A. Ealey and P. A. Davis, “Standard SELECT Electrostrictive Lead 57
R. W. Vest, “Metallo-Organic Decomposition (MOD) Processing of Fer-
Magnesium Niobate Actuators for Active and Adaptive Optical Components,” roelectrics and Electrooptic Films: A Review,” Ferroelectrics, 102, 53–68
Opt. Eng., 29 [11] 1373–82 (1990). (1990).
23
L. E. Cross, “Relaxor Ferroelectrics,” Ferroelectrics, 76, 241–67 (1987). 58
G. H. Haertling, “PLZT Thin Films Prepared from Acetate Precursors,”
24
J. Zhao, Q. M. Zhang, N. Kim, and T. Shrout, “Electromechanical Proper- Ferroelectrics, 116, 51–63 (1991).
ties of Relaxor Ferroelectric Lead Magnesium Niobate—Lead Titanate Ceram- 59
L. M. Sheppard, “Advances in Processing of Ferroelectric Thin Films,” Am.
ics,” Jpn. J. Appl. Phys., Part I, 34 [10] 5658–63 (1995). Ceram. Soc. Bull., 71 [1] 85–95 (1992).
25 60
M. P. Harmer and D. M. Smyth, “Nanostructure, Defect Chemistry, and R. W. Schwartz, T. J. Boyle, S. J. Lockwood, M. B. Sinclair, D. Dimos, and
Properties of Relaxor Ferroelectrics,” ONR Final Rept. No. N00014-82-K-0190, C. D. Buchheit, “Sol–Gel Processing PZT Thin Films: A Review of the State-
Lehigh University, Bethlehem, PA, Feb. 1992. of-the-Art and Process Optimization Strategies,” Integr. Ferroelectr., 7, 259–77
26
R. P. Brodeur, K. Gachigi, P. M. Pruna, and T. R. Shrout, “Ultra-High (1995).
61
Strain Ceramics with Multiple Field-Induced Phase Transitions,” J. Am. Ceram. O. Auciello, A. I. Kingon, S. B. Krupanidhi, R. Ramesh, M. de Keijser,
Soc., 77 [11] 3042–44 (1994). G. J. Dormans, S. K. Dey, P. V. Alluri, B. A. Tuttle, R. W. Schwartz, R. E.
27 Jones Jr., and S. B. Desu, “Electroceramic Thin Films, Part 1: Processing,” MRS
G. H. Haertling, “Electro-optic Ceramics and Devices”; pp. 371–492 in
Electronic Ceramics. Edited by L. M. Levinson. Marcel Dekker, New York, Bull., 21 [6] 25–58 (1996).
62
1988. Z. Y. Meng, U. Kumar, and L. E. Cross, “Electrostriction in Lead Lantha-
28 num Zirconate Titanate (PLZT) Ceramics,” in Piezoelectric and Electrostrictive
G. Goodman, “Ceramic Capacitor Materials”; pp. 79–138 in Ceramic Ma-
terials for Electronics. Edited by R. C. Buchanan. Marcel Dekker, New York, Materials for Transducer Applications, ONR Annual Rep. No. N00014-82-
1986. K0339, Pennsylvania State University, University Park, PA, May 1985.
29 63
R. Manczok and R. Wernicke, “Ceramic Boundary-Layer Capacitors,” H. Banno, “Piezoelectric Transducer and Piezoelectric Ceramics”; pp.
Philips Tech. Rev., 41 [11/12] 338–46 (1983). 2017–23 in Encyclopedia of Advanced Materials. Edited by R. W. Cahn. Per-
30 gamon Press, Oxford, U.K., 1995.
D. C. Hill and H. L. Tuller, “Ceramic Sensors: Theory and Practice”; pp.
64
265–374 in Ceramic Materials for Electronics. Edited by L. M. Levinson. Mar- V. A. Isupov, “Properties of Pb(Ti,Zr)O3 Piezoelectric Ceramics and Na-
cel Dekker, New York, 1988. ture of Their Orientational Dielectric Polarization,” Sov. Phys.–Solid State
31 (Engl. Transl.), 10 [4] 989–91 (1968).
B. Kulwicki, “PTCR Materials Technology, 1955–1980”; pp. 138–54 in
65
Advances in Ceramics, Vol. 1, Grain Boundary Phenomena in Electronic Ce- L. Benguigui, “Thermodynamic Theory of the Morphotropic Phase Tran-
ramics. Edited by L. M. Levinson. American Ceramic Society, Westerville, OH, sition Tetragonal–Rhombohedral in the Perovskite Ferroelectrics,” Solid State
1981. Commun., 11, 825–28 (1972).
32 66
G. Shirane, K. Suzuki, and A. Takeda, “Phase Transitions in Solid Solu- S. T. Liu, J. D. Heaps, and O. N. Tufte, “The Pyroelectric Properties of the
tions of Lead Zirconate and Lead Titanate: II,” J. Phys. Soc. Jpn., 7 [1] 12–18 Lanthanum-Doped Ferroelectric PLZT Ceramics,” Ferroelectrics, 3, 281–85
(1952). (1972).
33 67
B. Jaffe, R. S. Roth, and S. Marzullo, “Piezoelectric Properties of Lead R. R. Zeyfang, W. H. Sehr, and K. V. Kiehl, “Enhanced Pyroelectric Prop-
Zirconate–Lead Titanate Solid-Solution Ceramics,” J. Appl. Phys., 25 [6] 809– erties at a FE–FE Phase Transition,” Ferroelectrics, 11, 355–58 (1976).
68
10 (1954). R. W. Whatmore, P. C. Osbond, and N. M. Shorrocks, “Ferroelectric Ma-
34 terials for Thermal IR Detectors,” Ferroelectrics, 68, 351–67 (1987).
G. H. Haertling, “Piezoelectric and Electrooptic Ceramics”; pp. 135–225 in
818 Journal of the American Ceramic Society—Haertling Vol. 82, No. 4
69
W. A. Schulze and T. R. Gururaja, “Ferroelectric Ceramics for Dielectric Proceedings of the 10th IEEE International Symposium on Applications of
Electromechanical and Pyroelectric Applications,” MRS Bull., 12 [10/11] 48–52 Ferroelectrics (ISAF-96), Vol. 1 (New Brunswick, NJ, Aug. 1996). Edited by
(1987). B. M. Bulwicki, A. Amin, and A. Safari. IEEE, Piscataway, NJ, 1996.
70 81
K. Okazaki and K. Nagata, “Effects of Grain Size and Porosity on Electrical R. E. Newnham, “Composite Electroceramics,” CHEMTECH, Part 1, pp.
and Optical Properties of PLZT Ceramics,” J. Am. Ceram. Soc., 56 [2] 82–86 732–39, Dec. 1986; Part 2, pp. 38–45, Jan. 1987.
82
(1973). T. R. Gururaja, A. Safari, R. E. Newnham, and L. E. Cross, “Piezoelectric
71
G. H. Haertling, “Photoelectronic Effects in PLZT Ceramics”; pp. 21–50 in Ceramic–Polymer Composites for Transducer Applications”; pp. 92–128. in
Ceramic Transactions, Vol. 14, Electro-Optics and Nonlinear Optic Materials. Electronic Ceramics. Edited by L. M. Levinson. Marcel Dekker, New York,
Edited by A. Bhalla, E. Vogel, and K. Nair. American Ceramic Society, Wester- 1988.
83
ville, OH, 1990. L. E. Cross, “Ferroelectric Ceramics: Materials and Applications Issues”;
72
T. Takenaka, A. S. Bhalla, L. E. Cross, and K. Sakata, “Dielectric, Piezo- pp. 15–55 in Ceramic Transactions, Vol. 68, Hybrid Microelectronic Materials.
electric, and Pyroelectric Properties of Lead Zirconate–Lead Zinc Niobate Ce- Edited by K. M. Nair and V. N. Shukla. American Ceramic Society, Wester-
ramics,” J. Am. Ceram. Soc., 72 [6] 1016–23 (1989). ville, OH, 1996.
73 84
Y. Sugawara, K. Onitsuka, L. Yoshikawa, Q. Xu, R. E. Newnham, and K. T. R. Gururaja, “Piezoelectrics for Medical Ultrasonic Imaging,” Am. Ce-
Uchino, “Metal–Ceramic Composite Actuators,” J. Am. Ceram. Soc., 75 [4] ram. Soc. Bull., 73 [5] 50–55 (1994).
85
996–98 (1992). J. F. Scott, F. M. Ross, C. A. Araujo, M. C. Scott, M. Huffman, W. L.
74
A. Dogan, Q. Xu, K. Onitsuka, S. Yoshikawa, K. Uchino, and R. E. Newn- Warren, D. Dimos, R. M. Waser, A. I. Kingon, S. K. Streiffer, C. Basceri, S. R.
ham, “High-Displacement Ceramic–Metal Composite Actuators (Moonies),” Summerfelt, D. K. Fork, F. Armani-Leplingard, J. J. Kingston, D. L. Polla, and
Ferroelectrics, 156, 1–6 (1994). L. F. Francis, “Electroceramic Thin Films, Part II: Device Applications,” MRS
75
A. Dogan, J. Fernandez, K. Uchino, and R. E. Newnham, “Cymbal Elec- Bull., 21 [7] 33–65 (1996).
86
tromechanical Actuator”; pp. 213–16 in Proceedings of the 10th IEEE Interna- G. H. Haertling, “PZT and PLZT Thick Films on Silver, Sapphire, and
tional Symposium on Applications of Ferroelectrics (ISAF-96), Vol. 1 (New Silicon”; pp. 15–30 in Ceramic Transactions, Vol. 86, Intgrated Thin Films and
Brunswick, NJ, Aug. 1996). Edited by B. M. Kulwicki, A. Amin, and A. Safari. Applications. Edited by R. K. Pandy, P. Morris, A. Roshko, U. Varshney, and
IEEE, Piscataway, NJ, 1996. D. Witter. American Ceramic Society, Westerville, OH, 1998.
76 87
G. H. Haertling, “Rainbow Actuators and Sensors: A New Smart Technol- R. E. Newnham and G. R. Ruschau, “Smart Electroceramics,” Am. Ceram.
ogy,” Soc. Photo.-Opt. Instr. Eng. (SPIE), 3040, 81–92 (1997). Soc. Bull., 75 [10] 51–61 (1996).
77 88
P. S. Brody, “Optomechanical Bimorph Actuator,” Ferroelectrics, 50, 27– J. F. Scott, “Future Issues in Ferroelectric Miniaturization,” Ferroelectrics,
32 (1983). 206/207, 365–79 (1998).
78 89
K. Uchino, “Review: Photostriction and Its Application,” in Materials for M. Sayer, M. Lukacs, and T. Olding, “Emerging Technologies for Ferro-
Adaptive Structural Acoustic Control,” ONR Annual Rept. No. N00014-92-J- electric Films and Coatings,” Integr. Ferroelectr., 17, 1–10 (1997).
90
1510, Pennsylvania State University, University Park, PA, Mar. 1996. D. H. Lee, “Future Evolution of DRAM and Its Materials,” Integr. Ferro-
79
P. C. Lysne and C. M. Percival, “Analysis of Shock-Wave-Actuated Fer- electr., 17, 113–26 (1997).
91
roelectric Power Supplies,” Ferroelectrics, 10, 129–33 (1976). O. Auciello, J. F. Scott, and R. Ramesh, “The Physics of Ferroelectric
80
G. H. Haertling, “Stress-Induced Effects in PLZT Ceramics”; pp. 65–68 in Memories,” Phys. Today, [July] 22–27 (1998). 䊐

Gene Haertling received his B.S. degree in ceramic engineering from the University
of Missouri at Rolla in 1954. His M.S. and Ph.D. degrees, also in ceramic engineer-
ing, were earned from the University of Illinois in 1960 and 1961, respectively. From
1961 to 1973, he held staff and managerial positions at Sandia National Laboratories.
During this time he developed the first transparent ferroelectric ceramics, the PLZT
(lead lanthanum zirconate titanate) materials, which are now used in both military and
commercial applications. From 1974 to 1987, he was Vice-President of the Technical
Staff and Manager of the Ceramic Research Group at Motorola, Inc., Albuquerque,
NM. Just prior to joining Motorola, he was president of Optoceram, Inc., a small
entrepreneurial company he founded, which was engaged in the development and
manufacture of PLZT electrooptic ceramics. After briefly serving on the faculty at
University of Missouri at Rolla from 1987 to 1988, Dr. Haertling joined the Ceramic
Department of Clemson University as the Bishop Distinguished Professor of Ceramic
Processing. While at Clemson, he developed the special process for producing high-
displacement, piezoelectric ceramic actuators known as RAINBOWS. He is a mem-
ber of the National Academy of Engineers and a Fellow of The American Ceramic
Society and the IEEE. He has published 85 technical papers, 3 book chapters, and is
a coholder of 10 patents in the area of ferroelectric and electrooptic ceramic materials
and devices. Recently retired from active teaching, Dr. Haertling is Professor Emeri-
tus of Clemson University and is presently located in Albuquerque, NM.