Professional Documents
Culture Documents
Jiyan Dai
Author All books published by Wiley-VCH
are carefully produced. Nevertheless,
Prof. Jiyan Dai authors, editors, and publisher do not
The Hong Kong Polytechnic University warrant the information contained in
Department of Applied Physics these books, including this book, to
Hung Hom be free of errors. Readers are advised
Kowloon to keep in mind that statements, data,
Hong Kong illustrations, procedural details or other
items may inadvertently be inaccurate.
Cover
© antoniokhr/Getty Images; Library of Congress Card No.:
(graph) Courtesy of Professor Jiyan Dai, applied for
The Hong Kong Polytechnic Univeristy
British Library Cataloguing-in-Publication
Data
A catalogue record for this book is
available from the British Library.
10 9 8 7 6 5 4 3 2 1
v
Contents
2 Introduction to Ferroelectrics 15
2.1 What Is Ferroelectrics? 15
2.1.1 P–E Loop 15
2.1.2 Relationships Between Dielectric, Piezoelectric, Pyroelectric, and
Ferroelectric 16
2.1.2.1 Ferroelectric–Dielectric 16
2.1.2.2 Ferroelectric–Piezoelectric 17
2.1.2.3 Ferroelectric–Pyroelectric 18
2.2 Origin of Ferroelectrics 18
2.2.1 Structure-Induced Phase Change from Paraelectric to
Ferroelectric 18
2.2.2 Soft Phonon Mode 19
2.3 Theory of Ferroelectric Phase Transition 21
2.3.1 Landau Free Energy and Curie–Weiss Law 21
2.3.2 Landau Theory of First-Order Phase Transition 23
2.3.3 Landau Theory of a Second-Order Phase Transition 26
2.4 Ferroelectric Domains and Domain Switching 28
2.4.1 Domain Structure 28
2.4.2 Ferroelectric Switching 28
2.5 Ferroelectric Materials 29
vi Contents
4 Ferroelectric Characterizations 73
4.1 P–E Loop Measurement 73
4.2 Temperature-Dependent Dielectric Permittivity Measurement 76
4.3 Piezoresponse Force Microscopy (PFM) 77
4.3.1 Imaging Mechanism of PFM 77
4.3.2 Out-of-plane Polarization (OPP) and In-plane Polarization (IPP)
PFM 80
4.3.2.1 Electrostatic Force in PFM 83
4.3.2.2 Perspectives of PFM Technique 84
4.4 Structural Characterization 86
4.5 Domain Imaging and Polarization Mapping by Transmission Electron
Microscopy 87
4.5.1 Selected Area Electron Diffraction (SAED) 88
4.5.2 Convergent Beam Electron Diffraction (CBED) for Tetragonality
Measurement 91
References 92
10 217
Device Application of Multiferroics
10.1 ME Composite Devices 217
10.1.1 Effect of Preload Stress 221
10.2 Memory Devices Based on Multiferroic Thin Films 223
Contents ix
Index 247
1
Synapse
Nucleus
Axon
Axon
Memristor Input
Co
BFO Synapses
CCMO To pre-neuron
To post-neuron Output
Figure 1.2 A cross-bar structure of synapse and artificial neuron networks based on a
cross-bar structure. Ferroelectric thin films such as BiFeO3 can be used as the junction material.
Source: Adapted from Boyn et al. (2017).
Proportional
+ Integral Work
Sensor
Derivative
Actuator
(a) (b)
Figure 1.3 (a) Photo of a remote control copter and (b) diagram of a PID feedback control
system where sensors and actuators are implemented.
Y
MEMS scanner
X Horizontal scan
Z Vertical
scan
PZT on stainless steel
1 mm
Gyro sensors 4 mm
Resonant in Resonant in Resonant in
Y-direction Y-direction X-direction
Figure 1.4 FEM simulations of three resonant motions in a PZT-based gyroscope and photo of
a fabricated gyroscope. Source: Adapted from Chang and Chen (2017).
such as movement, electrical signals, thermal or magnetic energy, etc. The type
of input or output transducers being used really depends on the type of signal
or process being “sensed” or “controlled,” but we can define a transducer as a
device that converts one physical quantity into another.
A smart system needs sensors and actuators to realize the sensing functions
such as distance, movement, and acceleration as well as actions. These sensors
and actuators use smart materials to realize the conversion between different
energies and moduli to electrical signals such as voltage, current, and capaci-
tance. Of course, many sensor devices are made of semiconductors such as the
FET, but this is not the focus of this book.
“Smart material” is a very large concept, in fact, there is no stupid material (a
joke), i.e. all materials are smart in some way since they all have their own prop-
erties and response to external stimuli. But in this book, we restrict the “smart
materials” to those materials with “ferroic” characteristics. We focus on basic
physics, materials science, structures, devices, and applications of ferroic mate-
rials for smart systems. The ferroic materials are usually classified as possessing
one of the followings based on coupling of stimuli:
(i) Ferroelectric, which is also piezoelectric when electromechanically coupled
and pyroelectric when thermoelectrically coupled.
(ii) Ferromagnetic, which is also magnetostrictive when magnetomechanically
coupled.
(iii) Ferroelastic, which also includes shape memory when thermomechanically
coupled.
Among these ferroics, we can see that strain, electric polarization and magne-
tization, and their interplay or coupling are involved. We call a material as ferroic
material if it possesses at least one of the properties of ferroelectric, ferromag-
netic and ferroelastic.
If we look at the diagram shown in Figure 1.5, we can see that the coupling
and interplay between electricity, mechanics, magnetism, heat, and optics result
in many smart functions, such as ferroelectric, piezoelectric, pyroelectric, ferro-
magnetic, electromechanical, etc. One book cannot cover all of them, but those
belong to ferroic materials and devices especially in the form of thin films will be
1.2 Device Application of Ferroelectric Materials 5
Electricity
E
ici
ty Ma lectr
Elec
tr
ic
gn om
lec ion
alor
eto ag
oe ict ele ne
tro-o
ez str ctr tism
troc
Pi ctr
o
ici
e
ptic
El ty
Elec
Mechanics El Magnetism
as Magnetostriction
to-
op ric
tic
s c alo
to
optic
The
Elas oelastic
ne
ag
rm
M
toca
eto-
elec city
Opto tovoltai
tricit
Pho
i
n
The roelectr
loric
ism
Mag
elec
net
ag
rmo
tricit
om
Py
erm
y
cs
Th
Thermo-optics
Heat Optics
Figure 1.5 Diagram showing coupling between different moduli and the clarification of smart
materials.
extensively introduced in this book. Before going into details, some application
examples of ferroic materials in smart systems are given in this chapter.
(a) (b)
Figure 1.6 Piezoelectric materials-based sonar system for car (a) and submarine (b).
(a) (b)
Figure 1.7 (a) Transducers and (b) B-mode image of a wire phantom acquired with
PolyU-made array ultrasound transducer.
1.2 Device Application of Ferroelectric Materials 7
PMUT unit
Capping layer
Piezoelectric layer
Cavity
CMOS wafer
of fingerprint with a certain depth. This can overcome the problems of the cur-
rent fingerprint identification system in most mobile phones. InvenSense, Inc. is
one of the main suppliers of this solution, and Figure 1.8 is an illustration of the
ultrasonic fingerprint system.
Window
(a) (b)
Figure 1.9 A photo of an infrared detector (a) and illustration of its internal structure (b).
8 1 General Introduction: Smart Materials, Sensors, and Actuators
Spontaneous
polarization of
ferroelectric layer
Gate Memory
window
Source Drain – – – –
– – – – – – – –– –
N N
–––––––––
IDS P
–––––––––
– – – –
VG
Figure 1.10 Schematic diagram of field-effect transistor (FET) and the current–voltage (I–V)
characteristics induced by two different polarization state.
1.3 Device Application of Ferromagnetic Materials 9
Free layer
Pin layer
Figure 1.11 Schematic diagram of spin valve structure where arrows indicate the
magnetization directions.
materials are rare and are usually strong in one property but very weak in
another, such as BiFeO3 , which is very strong in ferroelectrics but very weak in
ferromagnetic (it is antiferromagnetic in fact). This makes multiferroic materials
hard to be practically applied in devices. However, people have been trying to
make composite materials such as piezoelectric with magnetostrictive mate-
rials, where the mechanical coupling between them makes the “multiferroic”
meaningful for device application, for example, making very sensitive magnetic
field sensor.
The magnetoelectric (ME) effect is the phenomenon of inducing magnetiza-
tion by an applied electric field (E) or polarization by magnetic field (H). Many
efforts have been devoted to improve the limit of detection of the ME compos-
ite at low frequency range, and values of ∼10−7 Oe at 1 Hz has been reported
(Wang et al. 2011). Based on the magnetic–strain–electric coupling, scientists
have demonstrated dc magnetic field sensor with a detection limit of 2 × 10−5 Oe
to dc magnetic field with a nonlinear ME magnetic effect (Li et al. 2017). Ferro-
electric material PZT and magnetostrictive material Metglas have been imple-
mented in the composite device (see Figure 1.12).
ID electrodes
Metglas
Piezofiber
Epoxy Kapton
VirginiaTech
Figure 1.12 Outline of ME device from Virginia Tech and schematic of the cross-section of the
ME composite. Source: Wang et al. (2011). Adapted with permission of John Wiley and Sons.
He
ati
ng
Stress/load
Loading
Cooling
Heating
Temperature
Figure 1.13 Phase transition between high temperature austenite phase and low
temperature martensite phase, where the shape at cubic-austenite phase can be resumed
from low temperature martensite phase whose lattice can be largely twisted.
into the jet flow based on the flight condition. As shown in Figure 1.14a,
SMAs activated by heat were developed that would allow for full chevron
immersion in jet flow during high thrust requirements (e.g. during take-off )
and not immersing it during cruise where the thrust efficiency is of greater
importance (Anon n.d.).
12 1 General Introduction: Smart Materials, Sensors, and Actuators
(a) (b)
Figure 1.14 (a) Brace of orthodontia using shape memory alloys and (b) arthrodesis device
developed by Karnes et al.
For broken bone rehabilitation, a SMA plate with a memory transfer tempera-
ture close to body temperature can be attached to both ends of the broken bone as
shown in Figure 1.14b. From body heat, the plate will contract and retain its origi-
nal shape, therefore exerting a compression force on the broken bone at the place
of fracture. After the bone has healed, the plate continues exerting the compres-
sive force and aids in strengthening during rehabilitation (Garlock et al. 2017).
References
Boyn, S., Grollier, J., Lecerf, G. et al. (2017). Learning through ferroelectric domain
dynamics in solid-state synapses. Nature Communications 8: 1–7.
Chang, C.-Y. and Chen, T.-L. (2017). Design, fabrication, and modeling of a novel
dual-axis control input PZT gyroscope. Sensors 17 (11): 2505.
Cho, J. (2018). Amid contradictory forecast: IC insights: ‘Memory chips will grow at
annual rate of 5% only on average by 2022’. Seoul, Korea: BusinessKorea.
Garlock, A., Karnes, W.M., Fonte, M. et al. (2017). Arthrodesis devices for
generating and applying compression within joints. US 2017/0296241 A1,
Available at: https://patents.google.com/patent/US20170296241A1/en.
Li, M., Dong, C., Zhou, H. et al. (2017). Highly sensitive DC magnetic field sensor
based on nonlinear ME effect. IEEE Sensors Letters 1 (6): 1–4.
Renesas Electronics Corporation (2017). Renesas electronics achieves large-scale
memory operation in fin-shaped MONOS flash memory for industry’s first
high-performance, highly reliable MCUs in 16/14nm process nodes and beyond.
Wang, Y., Gray, D., Berry, D. et al. (2011). An extremely low equivalent magnetic
noise magnetoelectric sensor. Advanced Materials 23 (35): 4111–4114. Available
at: https://doi.org/10.1002/adma.201100773.
15
Introduction to Ferroelectrics
E
–Ec +Ec
–Pr
Dipole
Electronic Ionic
reorientation
E=0
E>0
Electric field
Figure 2.2 Schematic diagram showing the origin of the electric polarization.
For all dielectric materials, the electron clouds deform under electric field,
forming electric dipoles. Electric polarization from this electron clouds defor-
mation is usually much smaller compared with ionic displacement. In ionic
crystals, when an electric field is applied, cations and anions are attracted to the
cathode and anode, respectively. While for a ferroelectric, electric polarization
reorientation (rotation or reversing) will result in a significant change of electric
displacement (see Figure 2.2).
This diagram also illustrates that a ferroelectric falls into a larger category called
dielectrics. Let’s look at their difference with an example by comparing refrac-
tive index of a typical dielectric material SiO2 and a typical ferroelectric mate-
rial BaTiO
√ 3 . By following relationship of refractive index with dielectric constant
n = 𝜖r , with dielectric constant of about 3.6 for SiO2 , its refractive index is
about 1.4–1.5. However, for a ferroelectric material, its dielectric constant is usu-
ally very large due to the reorientation of electric polarization, a few hundred for
BaTiO3 , for example, but we cannot
√ say that its refractive index is also very large
based on the relation of n = 𝜖r . In fact, the refractive index of a material is an
optical constant that only works in optical frequency regime in which electron
cloud can respond to the optical frequency of electric field (say 1014 Hz), but the
ferroelectric polarization
√ cannot follow the optical frequency to rotate or switch,
i.e. this equation n = 𝜖r does not work for ferroelectrics in optical frequency.
2.1.2.2 Ferroelectric–Piezoelectric
Many materials such as AlN, GaN, and ZnO exhibit electric polarization orig-
inated from their crystal structure symmetry breaking, but their polarizations
cannot be switched by external electric field. These electrically polarized mate-
rials are piezoelectric but not ferroelectrics. Figure 2.3 shows the relationships
between the piezoelectric and ferroelectric materials as well as pyroelectric and
dielectric materials. A ferroelectric is also piezoelectric since the electric polar-
ization change induced by mechanical strain will generate electric charges and
voltage at the surface of a ferroelectric plate, i.e. piezoelectricity.
18 2 Introduction to Ferroelectrics
Pyroelectric
Piezoelectric
Dielectric
2.1.2.3 Ferroelectric–Pyroelectric
As illustrated in Figure 2.3, a ferroelectric is also a pyroelectric, but a pyroelectric
material may not be a ferroelectric. Pyroelectricity is a phenomenon of tempera-
ture dependence of the spontaneous polarization. As the temperature of a crys-
tal is changed, electric charges corresponding to the change of the spontaneous
polarization appear on the surface of the sample. Among many pyroelectric mate-
rials, only those whose spontaneous polarization can be reversed by an electric
field belong to ferroelectrics.
Crystal system
Symmetry Polarity
Cubic Hexagonal Tetragonal Rhombohedral Orthorhombic Monclinic Triclinic
Centro (11) m3m m3 6/mmm 6/m 4/mmm 4/m 3m 3 mmm 2/m
32 point group Non-polar
432 622 422
of (22) 23 6 4 32 222
crystallographic Non-centro 43m 6m2 42m
symmetry (22) 2
Polar (10) 6mm 6 4mm 4 3m 3 mm2 1
m
Ti
Ti O Ti
Ps
Ps
Ba
T > Tc T < Tc
• Among these 20 point groups, 10 of them have a unique polar direction, and
these 10 polar classes are pyroelectrics. Among these pyroelectrics, only those
whose polar can be reversed by external electric field are ferroelectrics.
BaTiO3 is an example of a typical ferroelectric crystal with perovskite struc-
ture as illustrated in Figure 2.4. Above the Curie temperature T c , the paraelectric
cubic structure is centrosymmetric, whereas in the tetragonal phase below T c , it
is energetically favorable for the O2− ions to be shifted slightly below face centers,
and Ti4+ ions are shifted upward from the unit cell center. The relative change in
positions of the Ti4+ and O2− ions produces a spontaneous polarization Ps as well
as the non-centrosymmetric structure. More ferroelectric phases of BaTiO3 will
be introduced later.
Elongation
direction
(a) (b)
(c)
Figure 2.5 Starting from the original cubic structure (a), if (b) is stabilized, only oxygen
octahedra are distorted without generating dipole moments (acoustic mode). When (c) is
stabilized, dipole moments are generated (optical mode). The final stabilized state (c)
corresponds to ferroelectric state.
Since only oxygen octahedra are distorted in Figure 2.5b without generating
electric polarization, its vibration is a kind of phonon mode. However, if the par-
ticular mode Figure 2.5c becomes stabilized, with decreasing temperature, the
vibration frequency decreases (soft phonon mode), and finally at a certain phase
transition temperature (Curie temperature), this frequency becomes zero, i.e. the
electric polarization state is stabilized.
Since the soft modes in ferroelectrics lead to electric polarization, they are opti-
cally active and can be detected by means of optical spectroscopy in the spectra of
dielectric permittivity (real and imaginary parts). Therefore, the vibration shown
in Figure 2.5c is also called optical mode.
Spectroscopic studies of the soft phonon modes provide a very powerful tool
for investigating the ferroelectric transitions. Figure 2.6 shows Raman spectra of
BaTiO3 phase transition, where the 305 cm−1 peak reduces its sharpness as the
in situ temperature increases and becomes indistinct at 150 ∘ C, indicating that
tetragonal phase is transformed to cubic phase. This trend is consistent with the
results that the peak around 310 cm−1 appeared below Curie temperature (T c )
and vanished above T c (Hayashi et al. 2013).
The dielectric permittivity in the soft mode is governed by three laws (Cochran
1960; Shirane et al. 1970; Luspin et al. 1980):
(a) Static dielectric permittivity 𝜖 r (0) produced by the soft mode obeys the
Curie–Weiss law:
(b) The eigen frequency 𝜔(T) of the soft mode follows the Cochran behavior:
1
𝜔(T) ∼ (T − T0 ) 2 (2.7)
Intensity (a.u.)
200 °C
150 °C
100 °C
50 °C
25 °C
Figure 2.6 In situ Raman spectra of BaTiO3 particles measured at different temperatures: 25,
50, 100, 150, and 200 ∘ C. Source: Hayashi et al. (2013). Adapted with permission of Elsevier.
(c) The static dielectric constant and the soft mode frequency are connected via
the Lyddane–Sachs–Teller (LST) relation:
𝜖∞ 𝜔2 (T)
(T) = T 2 (2.8)
𝜖r (0) 𝜔L
where 𝜖 ∞ is the high-frequency dielectric constant and 𝜔T and 𝜔L are the
transverse and longitudinal frequencies of the corresponding vibrations,
respectively.
where F 0 does not depend on the order parameter P and it describes the tem-
perature dependence of free energy of high temperature phase near the phase
transition (Graz University of Technology n.d.; Uchino 2009). The coefficients
𝛼, 𝛽, and 𝛾 are temperature dependent in general, while their signs are different
for the first- and second-order phase transitions. The reason to have only even
number of powers is that when the polarization changes signs, the free energy of
the crystal will not change. Since this phenomenological formula describes the
free energy across the paraelectric to ferroelectric phase transition, it should be
applicable to all temperature ranges.
In our calculation, however, the coefficient 𝛼 is the only coefficient that is
assumed to be temperature dependent. The fraction numbers are for conve-
nience when we differentiate F in respect of P. To define the relation of 𝛼 and T,
the free energy for certain polarization should be considered. As in paraelectric
state (i.e. P = 0), the free energy should be zero (F(P, T) = 0) at any temperature
above its Curie point. In ferroelectric state, there are two situations for free
energy, either F(P, T) > 0 or F(P, T) < 0. For free energy greater than zero, the
paraelectric state should be realized because it always tends to the minimum
energy state. Therefore, the free energy for a certain polarization must be lower
than zero in order to stabilize the ferroelectric state. Thus, the coefficient 𝛼 of
the P2 term must be negative in ferroelectric state, while it is positive passing
through zero at temperature T 0 . According to this concept, a relation of 𝛼 and T
is formed as the following:
𝛼 = 𝛼0 (T − T0 ) (2.11)
where 𝛼 0 is a temperature-independent constant and T 0 the Curie–Weiss tem-
perature, which is equal to or lower than the actual transition temperature T c
(Curie temperature).
By differentiating the free energy, electric field E related to the equilibrium
polarization can be expressed as
𝜕F
E= = 𝛼P + 𝛽P3 + 𝛾P5 (2.12)
𝜕P
By considering the existence of polarization without external electric field, i.e.
E = 0, we can get
P(𝛼 + 𝛽P2 + 𝛾P4 ) = 0 (2.13)
This equation has two possible solutions, among them the trivial solution P = 0
that corresponds to a paraelectric
√
state is not our concern for ferroelectric state.
−𝛽± 𝛽 2 −4𝛼𝛾
Another finite solution P2 = 2𝛾
corresponds to a ferroelectric state.
From Landau free energy, the temperature-dependent dielectric constant
(relative permittivity) of a ferroelectric material at ferroelectric and para-
electric phases (below and above Curie temperature) can be derived. For
the first order of phase transition from paraelectric phase to ferroelectric
phase, the temperature-dependent Landau free energy, dielectric constant, and
spontaneous polarization are illustrated in Figure 2.7. One can see that the
characteristic of the first-order ferroelectric phase transition is that the order
parameter Ps drops to zero abruptly and the dielectric constant reaches a finite
2.3 Theory of Ferroelectric Phase Transition 23
T > T1
T c < T < T1
1/ϵr
Ps
T = Tc
T0 < T < Tc ϵr
T < T0
(a) (b) T0 Tc T1
Figure 2.7 (a) Free energy, (b) dielectric constant and spontaneous polarization, of first-order
phase transition.
√ √
−𝛽 + 𝛽 2 − 4𝛼0 (T − T0 )𝛾
P = 0, ± for T0 < T < T1
2𝛾
P=0 for T > T1
The temperature T 1 is where the two nonzero solutions become unstable. For
T > T 1 , there is only one free energy minimum. For T 1 > T > T 0 , there are three
√ minima. P = 0 is the lowest free energy solution when T 1 > T > T c , while
potential
√
−𝛽+ 𝛽 2 −4𝛼0 (T−Tc )𝛾
P=± 2𝛾
exhibits the lowest free energy when T c > T > T 0 . For
T < T 0 , there are double potential minima of the free energy that correspond to
stable spontaneous polarization.
The difference between the Curie–Weiss temperature T 0 , the Curie tempera-
ture T c , and the ferroelectric limit temperature T 1 can be verified based on the
potential minima obtained in the first-order phase transition.
As the potential minima are obtained from
𝜕F
= E = 𝛼P + 𝛽P3 + 𝛾P5 = 0
𝜕P
which is valid for any temperature below or above Curie temperature, there are
three possible minima including P = 0 (i.e. F = 0). At T = T c , the free energy at
nonzero polarization must be equal to that of the paraelectric state, i.e.
F(P, T)|T=Tc = F(P, T)|P=0 = 0
So,
1 2 1 4 1 6
F(P, T) = 𝛼P + 𝛽P + 𝛾P = 0
2 4 6
then we get
𝛼 + 𝛽P2 + 𝛾P4 = 0
1 1
𝛼 + 𝛽P2 + 𝛾P4 = 0
2 3
By eliminating the P4 term from these two equations,
1 2
2𝛼 + 𝛽P = 0
2
4𝛼
P2 = −
𝛽
Putting P2 = − 4𝛼
𝛽
into the equations,
( ) ( )2
4𝛼 4𝛼
𝛼+𝛽 − +𝛾 − =0
𝛽 𝛽
or
( )2
1 4𝛼
𝛼− 𝛾 − =0
3 𝛽
2.3 Theory of Ferroelectric Phase Transition 25
Taking 𝛼 = 𝛼 0 (T − T 0 ) = 𝛼 0 (T c − T 0 ),
3 𝛽2
𝛼 = 𝛼0 (Tc − T0 ) =
16 𝛾
the Curie temperature is thus calculated as
3 𝛽2
Tc = T0 +
16 𝛼0 𝛾
This indicates that T c is little higher than T 0 . Meanwhile, when T = T 1 , the free
energy has only one solution at P = 0, i.e.
𝛽 2 − 4𝛼0 (T1 − T0 )𝛾 = 0
𝛽2
T1 = T0 +
4𝛼0 𝛾
For this equation, we can identify that T 1 is higher than T c .
As a result, we can conclude that
T0 < Tc < T1
To identify the dielectric constant, we first have to calculate the relative permit-
tivity 𝜖 r which is the response of the system to an electric field.
dP
𝜖r =
dE
In free energy equation, the field has been included to calculate 𝜖 r
𝜕F
=E
𝜕P
E = 𝛼0 (T − Tc )P + 𝛽P3 + 𝛾P5
dP || 1
at Tc , 𝜖r = =
dE ||P=0 𝛼0 (T − Tc )
dP || 1
at T1 , 𝜖r = | √ −𝛽 =
dE |P= , 𝛼0 (T − T1 )
2𝛾
Based on the equations earlier, we can verify that in case of 𝛽 < 0, the permittivity
shows a maximum and a discontinuity of the spontaneous polarization appears
at T c .
26 2 Introduction to Ferroelectrics
T > Tc
T = Tc
Ps
1/ϵr
T < Tc ϵr
(a) (b) Tc
Figure 2.8 (a) Free energy, (b) dielectric constant and spontaneous polarization, of
second-order phase transition.
2.3 Theory of Ferroelectric Phase Transition 27
dP || 1 C
𝜖r = = = for T > Tc
dE ||P=0 𝛼0 (T − Tc ) (T − Tc )
dP || 1 C
𝜖r = = = for T < Tc
dE ||P2 = 𝛼0 (Tc − T) 2𝛼0 (Tc − T) 2(Tc − T)
𝛽
where C = 𝛼1 is Curie constant.
0
Now we can see that the dielectric permittivity can be derived from
second-order differentiation of Landau free energy with boundary conditions of
E = 0 (without electric field). The temperature-dependent dielectric permittivity
for ferroelectric materials following the first-order phase transition can thus be
derived. In both cases, 1∕𝜖r linearly depends on temperature.
The classification of order of phase transition comes from thermodynamic free
energy according to Ehrenfest (Jaeger 1998; Blundell and Blundell 2009), where
it can be labeled as the lowest derivation (order parameter) of the free energy
that shows discontinuity as a function of other thermodynamic variables during
phase transition. For instance, if a discontinuity is exhibited in the first derivative
of Gibbs free energy, it is classified as first-order phase transition. If the discon-
tinues order parameter is the second order of differentiation of Gibbs free energy,
the corresponding phase transition is classified as second-order phase transi-
tion. Although Ehrenfest classification is clear to understand, it is not a complete
method as it is not suitable if the derivative of free energy diverges. Therefore,
a modern classification is made. Similar to the Ehrenfest classes, two broad cat-
egories are divided by taking the latent heat into account (Maris and Kadanoff
1978).
For example, various solid/liquid/gas transitions usually involve a discontinu-
ous change in density, which is the first derivative of free energy with respect to
pressure. Therefore, they are classified as first-order phase transition based on
Ehrenfest classification. In the view point of modern classification, during the
water to ice phase transition, latent heat is added while the temperature of the
system remains unchanged. The specific discontinuity with respect to energy and
temperature of this phase transition thus assorts to the first-order differentiation
of Gibbs free energy, and it belongs to first-order phase transition associated with
latent heat. While for β-brass (an alloy of copper and zinc, in equal amounts)
phase transition, there is no latent heat. However, when the temperature is lower
than transition temperature, the probability for each copper atom to have more
zinc nearest neighbors increases, which would result in a completely random
arrangement of copper and zinc atoms. To say, the material undergoes a contin-
uous phase transition in which the specific heat has a singularity and it belongs
to second-order phase transition.
For the ferroelectric phase transition, for example, BaTiO3 , there is latent
heat during paraelectric to ferroelectric phase transition, and the spontaneous
polarization abruptly (discontinuously) changes to zero, which is the first-order
differentiation of Landau free energy. Therefore, this phase transition belongs
to first-order phase transition. While for the second-order phase transition, the
Ps continuously changes to zero, but the dielectric permittivity, which is the
second-order of differentiation of Landau free energy, is discontinuous at T c .
28 2 Introduction to Ferroelectrics
Most ferroelectric phase transitions are first order, very few (such as triglycine
sulfate [TGS]) are second-order transition.
β+
γ– α–
α+ γ+
(a) (b) β–
Figure 2.9 (a) Ferroelectric domain structure in BaTiO3 and (b) cloverleaf domain patterns and
the enlarged single vortex domain in the ferroelectric YMnO3 . Source: Adapted from Choi et al.
(2010) and Reprinted with permission of Springer Nature: Zhang et al. (2013).
2.5 Ferroelectric Materials 29
E-field
direction E-field
direction
Stress
90° switching
Figure 2.10 Illustration of polarization switching by electric field and mechanical stress.
as well as up-to-down, i.e. 90∘ and 180∘ polarization rotations, respectively. How-
ever, a mechanical stress along the long axis can only press the ellipsoid into an
in-plane ellipsoid.
In a ferroelectric crystal, its polarization switching is accompanied by domain
switching as well as domain wall movement. This process will induce strain in the
crystal, where the non-180∘ switching results in large piezoelectric strain.
a′
Ti
b′ 130 °C c
a′ –90 °C 0 °C
a
b
Ba
Rhombohedral Orthorhombic Tetragonal Cubic
(a = b = c; (a = b = c; (a = b ≠ c; (a = b = c;
α = β = γ ≠ 90°) α ≠ 90°, β = γ = 90°) α = β = γ = 90°) α = β = γ = 90°)
cubic structure to index the atomic planes and directions for convenience, but in
fact, the crystal structure is not cubic any more due to its distortion.
You may be wondering, the so-called orthorhombic structure in Figure 2.11
should be a monoclinic structure when the cubic unit cell is stretched along
[110] direction. It seems to be right, but since there is relative shift of cations
and anions, the real Bravais unit cell is actually a larger orthorhombic unit cell
with the shadow plane as the a′ b′ plane with c′ perpendicular
√ to a′ b′ plane.
′ ′ ′
The orthorhombic unit cell has a = a, b = c = 2a, i.e. the volume of the
orthorhombic unit cell is doubled, and a′ , b′ , and c′ are not equal, where a is for
pseudocubic unite cell in the distorted structure.
Ceramics based on BaTiO3 are very important in electronic component of
capacitors due to its high dielectric constant and low dielectric loss. BaTiO3 thin
film is also one of the most attractive ferroelectric thin films for memory device
application. This will be introduced in Chapter 4.
Pure SrTiO3 is a quantum paraelectric material where quantum fluctuations of
atomic positions suppress a ferroelectric transition, leading to a so-called incipi-
ent ferroelectric (Müller and Burkard 1979; Zhong and Vanderbilt 1996; Barrett
1952; Hemberger et al. 1995). It means that SrTiO3 is supposed to be ferroelectric
but its Curie temperature is absolute 0 K, which can never be reached. There-
fore temperature-dependent dielectric constant of SrTiO3 increases continuously
when temperature decreases, but it can never have a peak like a normal ferroelec-
tric material (see Figure 2.12). However, one can see that when a bias voltage is
applied on the SrTiO3 crystal, a relaxor-like paraelectric–ferroelectric transition
happens above 0 K. This is because of the existence of polar nano regions (PNRs)
in SrTiO3 at low temperatures. These PNRs can form ferroelectric nano-domains
under electric field, giving rise to relaxor-like ferroelectricity that will be intro-
duced later in this section.
SrTiO3 single crystal has been used as many oxide thin films substrate material
due to its high chemical stability and good lattice matching with many perovskite
oxide materials. When SrTiO3 is doped with BaTiO3 , it becomes (Bax Sr1−x )TiO3
(called BST) and its Curie temperature can be tuned between 0 K to T c of BaTiO3
depending on how much BaTiO3 is added. The study of BST, especially BST thin
films, has been a very hot topic due to its large dielectric tunability which is very
useful in tunable microwave device.
BST is attractive for microwave device application mainly because of two rea-
sons. The first is that the dielectric permittivity of BST strongly depends on the
2.5 Ferroelectric Materials 31
25
1
1- E = 0 kV/cm
20 2- E = 2 kV/cm
3- E = 5 kV/cm
4- E = 15 kV/cm
ϵ(T ) × 103
15
10
2
5 3
4
0
0 100 200 300
Temperature (K)
Figure 2.12 Dielectric constant of SrTiO3 single crystal as a function of temperature and
biasing field. Source: Vendik et al. (1999). Reprinted with permission of Springer Nature.
bias electric field; this property is termed as dielectric tunability. Another reason
is that BST possesses very low dielectric loss (tan 𝛿) at microwave frequency. In
the following paragraphs, properties and applications of BST are introduced.
In Figure 2.13, an experimental result shows the change in lattice parameters
when different compositions of Ba and Sr are added into the compound. When
the composition of Ba increases, it can be seen that the structure of BST varies,
changing from cubic to tetragonal, proofing the structural transition from SrTiO3
(STO) to BaTiO3 (BTO). Therefore, it can be concluded that the tetragonality
increases when the composition of Ba increases.
0.405
c
Lattice parameters (nm)
0.400
a T
0.395
C
0.390
Figure 2.13 Lattice constant of (Bax Sr1−x )TiO3 (both bulk and films) as a function of
composition x. Source: Adapted from Abe and Komatsu (1995).
32 2 Introduction to Ferroelectrics
25 000
x=0
20 000 x = 0.34
x = 0.27
x = 0.47 x = 0.65
x = 0.87
15 000
ϵr(T )
10 000 x = 1.0
5000
0
0 100 200 300 400
Temperature (K)
Figure 2.14 The phase transition behavior (the dielectric constant as a function of
temperature) of (Bax Sr1−x )TiO3 single crystal for various content of Ba–x. Source: Vendik et al.
(1999). Reprinted with permission of Springer Nature.
Figure 2.14 shows a theoretical result that the composition of Ba in BST varies
the Curie temperature as well as the dielectric properties of the compound.
Owing to the increase of tetragonality, when more Ba is added, the Curie
temperature increases. One can also notice that the dielectric constant near T c
is a few orders of magnitude greater than the rest, indicating a paraelectric to
ferroelectric phase transition.
Figure 2.15 shows the BST dielectric constant as a function of bias electric field.
The dielectric constant has a huge difference when E-field is applied. In the case
of x = 0.24, the dielectric constant is greatly reduced from 𝜖 r > 1400 to 𝜖 r < 500
when a field of E = 10 MV/m is applied. The electric field-dependent dielectric
permittivity is commonly described as the dielectric tunability n, which is defined
as the ratio of the dielectric permittivity at zero electric field bias to the permit-
tivity under electric field bias E, i.e.
𝜖r (0) − 𝜖r (E)
n=
𝜖r (0)
Question to students: Why does dielectric constant decrease when an E-field is
applied?
Dielectric constant ϵ
1000
800
0.24
600
0.44
400
0
0.68
200
1.0
0
–50 –40 –30 –20 –10 0 10 20 30 40 50
Bias field (MV/m)
Pb
Ti
Ti O
Ps
Figure 2.16 (a) Perovskite structure of PbTiO3 in the cubic form above T c and (b) tetragonal
structure of PbTiO3 for T < T c presenting spontaneous polarization.
the transition are named antiferroelectrics, and PbZrO3 remains the prototype
for this large group of materials.
When an antiferroelectric crystal is subject to electric fields, the antiparal-
lel dipoles can be flipped and forced to be parallel, corresponding to an elec-
tric field-induced antiferroelectric-to-ferroelectric phase transition (Cross 1967).
This phase transition manifests itself by the signature of double-hysteresis loops
as the electric polarization P is monitored as a function of applied field E (see
Figure 2.17a). In addition to the abrupt change in polarization, significant changes
in other quantities, such as linear dimensions (Berlincourt et al. 1964; Pan et al.
1989a; Brodeur et al. 1994) and optical properties (Chandani et al. 1989), also
accompany the phase transition. These changes at the phase transition provide
the physical foundation for important engineering applications of antiferroelec-
tric materials in digital displacement transducers (Berlincourt 1966; Uchino 1985;
Pan et al. 1989b), energy storage capacitors, electrocaloric cooling devices (Tut-
tle and Payne 1981; Mischenko et al. 2006; Scott 2011), and flat panel displays
(Fukuda et al. 1994; Takezoe et al. 2010). For most device applications, antiferro-
electric materials in the thin film form are required. As a result, films with various
antiferroelectric compositions have been deposited and characterized (Li et al.
1994; Qu and Tan 2006).
As illustrated in Figure 2.17b, with respect to the pseudocubic perovskite
primitive unit cell that contains
√ one PbZrO
√ 3 molecule, the dimensions of the
orthorhombic cell are a = 2ac , b = 2 2ac , and c = 2ac , where the parameter
ac is the lattice constant for pseudocubic unit cell.
The primary structural feature that accounts for the AFE is the antiparallel
displacements of Pb2+ along the pseudocubic ⟨110⟩ directions, as schematically
shown in Figure 2.17b. At room temperature, the average displacement of Pb2+
is ∼0.025 nm.
Energy storage is an important application of antiferroelectric material. As
illustrated in Figure 2.18a, the energy density W of dielectric capacitors is
4
E b
4
a
c
(a) (b)
Figure 2.17 (a) Typical P–E loop of antiferroelectric and (b) structure models for PbZrO3 -based
antiferroelectric. The orthorhombic unit cell is indicated by the box, while arrows represent the
direction of shifts of Pb ions along the c-axis. Source: Tan et al. (2011). Reprinted with
permission of John Wiley and Sons.
2.5 Ferroelectric Materials 35
D D D
E E E
Figure 2.18 The typical electric displacement verses electric field (D–E) hysteresis loops and
energy storage characteristics of the three classes of solid dielectric materials (a) linear,
(b) ferroelectric, and (c) antiferroelectric.
represented by the shadow area that can be calculated by the following formula:
Pmax
U= E ⋅ dD = E dD (upon discharging)
∮ ∫Pr
where Pr is remnant polarization and Pmax is the maximum polarization under
bias electric field E whose change generates electric displacement dD. We can see
that the linear dielectric has relatively low energy density (∼0.01 J/cm3 ) because of
the low dielectric constant. The ferroelectric material owns large electric polar-
ization and hence has very high dielectric permittivity. However, when electric
fields excess the coercive field of the material, the permittivity decreases very
rapidly with the electric fields. Moreover, because of the hysteresis characteristic
of ferroelectric materials, the energy stored in ferroelectric during the charging
process cannot be fully released (see Figure 2.18b). Instead, some of the energy is
converted into heat due to the internal friction of dipole moment reorientation.
The shape of the double-hysteresis loop of antiferroelectric, as shown in
Figure 2.18c, can benefit to the energy storage in two aspects. First, because
of the AFE–FE phase transition, AFE can sustain a high dielectric constant
in relatively high electric field and hence absorbing a large amount of energy.
On the other hand, ferroelectric structures can turn back to AFE phase at low
electric field and release nearly all the energy.
800
700 Cubic
600 20
–31
–32
Temperature (K)
500 –25
Rhombohedral(HT) –24
–25 19
MB 13
400 Tetragonal
MA
30
300
–25 –26 –22 –12 10 15 26 31
200
Rhombohedral(LT) 23
100 R3c/R3m+Cm
Orthorombic 16 19 22 30 R3m+P4mm+Cm
35
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Ti concentration
Figure 2.19 PZT phase diagram. The crossover between MB and MA structures is marked by a
dashed line. Tetragonal and rhombohedral phase regions are ferroelectric phases. Source:
Adapted from Zhang et al. (2014).
phases. That’s the point! One should notice that the best piezoelectric property
for PZT actually appears at the composition close to the MPB; one of the main
reasons for PZT to possess excellent piezoelectric property is the existence of
the MPB at which the monoclinic phase is the result of structural uncertainty
between the left side rhombohedra and the right side tetragonal structures. It
is because of this “uncertainty,” under a small applied electric field, the structure
distortion can easily happen between the tetragonal and rhombohedra phases, i.e.
non-180∘ polarization rotation, which is the key factor to induce very large strain
in the crystal. Of course, the large strain also comes from domain wall movement
that will be further discussed later.
Dielectric constant
ferroelectrics.
15 000
1000 f
30
Electric displacement (μC/cm2)
20
10
0
24 °C
–10 46 °C
72 °C
96 °C
–20
124 °C
152 °C
–30 165 °C
–10 –8 –6 –4 –2 0 2 4 6 8 10
(b) Electric field (kV/cm)
Lead magnesium niobate, Pb(Mg1/3 Nb2/3 )O3 (PMN), is one of the most
interesting relaxor dielectric materials due to its quenched order–disorder
complex structure. The ordered and disordered distribution of Mg2+ and Nb5+
cations on the equivalent lattice B sites in PMN is the essential structural
characteristic of relaxor dielectrics. While lead titanate, PbTiO3 (PT), is a normal
ferroelectric material with long-range order domain structure and has a sharp
phase transition taking place at the Curie temperature T c = 490 ∘ C. PMN-PT
is the solid solution of Pb(Mg1/3 Nb2/3 )O3 and PbTiO3 , in which the substitution
of Ti4+ cations in B sites tends to reduce the random field and enhances the
size and formation of long-range ordered ferroelectric domains. PMN-PT has a
complex perovskite structure with an ABO3 -type unit cell and exhibits a relaxor
ferroelectric property.
Micro-polar clusters interact Micro-polar develop but Regions are nonpolar and
and material exhibits a have random orientation due material exhibits no
remanent polarization to thermal excitation macroscopic polarization
Increasing temperature
Figure 2.21 Polar mechanisms in relaxor ferroelectric materials as postulated by Cross (1987).
In the superparaelectric phase, the materials exhibit a nonzero root mean square (RMS) dipole
moment but have zero macroscopic polarization. Source: Cross (1987). Reproduced with
permission of Taylor & Francis.
2.5 Ferroelectric Materials 39
the crystal. Although the information about PNRs can be indirectly measured by
neutron scattering, X-ray diffraction, high-resolution electron microscopy, etc.,
their evolution under different conditions cannot be directly observed and is still
unclear.
The PNRs with sizes around 1.5 and 18 nm in the Pb(Mg1/3 Nb2/3 )O3 and
Pb(Zn1/3 Nb2/3 )O3 crystals, respectively, at room temperature have been
reported (Gehring et al. 2001; Stock et al. 2004; Xu et al. 2004). Compared with
the size of PNRs in PMN determined by means of high-resolution electron
microscopy (Yoshida et al. 1998), PNRs size determined from elastic neutron
scattering is an order of magnitude smaller. Figure 2.22 shows piezoresponse
force (PFM) images of various compositions (001)-oriented PMN-xPT crystals
with different PT content x. It is apparent that the domain size reduces when the
content of PT is reduced, and when x is equal to 20%, the nanosized domains
dominant the domain structure.
The domain structure can also be observed under scanning electron micro-
scope (SEM) if the sample is treated by acid. Figure 2.23 is the laminate domain
structures of PMN-0.35PT crystal, where one can see that when the content of PT
2 μm
5 μm
40 2 Introduction to Ferroelectrics
increases to more than 30% (x > 0.3), laminate micro-sized domains are formed
to make the crystal more like a normal ferroelectric.
Pb(Mg1/3Nb2/3)TixO3
500
Cubic
400 Tm
Tetragonal
T (K)
R-O
300
M
Rhombohedral
200
0.1 0.2 0.3 0.4 0.5
X
Figure 2.24 Phase diagram of PMN-xPT. Source: Guo et al. (2002). Adapted with permission of
AIP Publishing.
2.6 Ferroelectric Domain and Phase Field Calculation 41
Electric field
Electric field
Figure 2.26 Schematic illustration of the mechanism of the reversible domain switching. The
light and dark gray regions represent the ferroelectric domains having tetragonal crystal
structure with deferent orientations, respectively. The black arrows indicate the direction of
the c-axis of the tetragonal phase. The small white rectangles represent the spatial symmetry
of the point defects (dopant/impurity ions and oxygen-ion vacancies) in ferroelectrics. E appel is
an electric potential energy induced from the external electric field. Adapted from Koyama
and Onodera (2009).
42 2 Introduction to Ferroelectrics
Electric field
50 μm 50 μm
References
Abe, K. and Komatsu, S. (1995). Ferroelectric properties in epitaxially grown
Bax Sr1−x TiO3 thin films. Journal of Applied Physics 77 (1995): 6461.
Bai, F., Li, J., and Viehland, D. (2004). Domain hierarchy in annealed (001)-oriented
Pb(Mg1/3 Nb2/3 )O3−x %PbTiO3 single crystals. Applied Physics Letters 85 (12):
2313–2315. Available at: https://doi.org/10.1063/1.1793353.
References 43
Barrett, J.H. (1952). Dielectric constant in perovskite type crystals. Physical Review
86 (1): 118–120.
Berlincourt, D. (1966). Transducers using forced transitions between ferroelectric
and antiferroelectric states. IEEE Transactions on Sonics and Ultrasonics 13 (4):
116–124.
Berlincourt, D., Krueger, H.H.A., and Jaffe, B. (1964). Stability of phases in modified
lead zirconate with variation in pressure, electric field, temperature and
composition. Journal of Physics and Chemistry of Solids 25 (7): 659–674. Available
at: http://www.sciencedirect.com/science/article/pii/0022369764901751.
Blundell, S.J. and Blundell, K.M. (2009). Concepts in Thermal Physics, 2e. Oxford:
OUP.
Brodeur, R.P., Gachigi, K., Pruna, P.M., and Shrout, T.R. (1994). Ultra-high strain
ceramics with multiple field-induced phase transitions. Journal of the American
Ceramic Society 77 (11): 3042–3044. Available at: https://doi.org/10.1111/j.1151-
2916.1994.tb04546.x.
Chandani, A.D.L., Gorecka, E., Ouchi, Y. et al. (1989). Antiferroelectric chiral
smectic phases responsible for the tristable switching in MHPOBC. Japanese
Journal of Applied Physics 28 (7): L1265–L1268.
Choi, T., Horibe, Y., Cheong, S.W. et al. (2010). Insulating interlocked ferroelectric
and structural antiphase domain walls in multiferroic YMnO3 . Nature Materials
9: 253. Available at: http://dx.doi.org/10.1038/nmat2632.
Cochran, W. (1960). Crystal stability and the theory of ferroelectricity. Advances in
Physics 9 (36): 387–423. Available at: https://doi.org/10.1080/
00018736000101229.
Cross, L.E. (1967). Antiferroelectric-ferroelectric switching in a simple ‘Kittel’
antiferroelectric. Journal of the Physical Society of Japan 23 (1): 77–82.
Cross, L.E. (1987). Relaxor ferroelectrics. Ferroelectrics 76 (1): 241–267. Available at:
https://doi.org/10.1080/00150198708016945.
Fu, H. and Cohen, R.E. (2000). Polarization rotation mechanism for ultrahigh
electromechanical response in single-crystal piezoelectrics. Nature 403: 281–283.
Available at: http://dx.doi.org/10.1038/35002022.
Fukuda, A., Takanishi, Y., Isozaki, T. et al. (1994). Antiferroelectric chiral smectic
liquid crystals. Journal of Materials Chemistry 4 (7): 997–1016. Available at:
http://dx.doi.org/10.1039/JM9940400997.
Gehring, P.M., Wakimoto, S., Ye, Z.G., and Shirane, G. (2001). Soft mode dynamics
above and below the Burns temperature in the relaxor Pb(Mg1/3 Nb2/3 )O3 .
Physical Review Letters 87 (27): 277601.
Graz University of Technology (n.d.). Landau theory of a first order phase transition.
Available at: http://lampx.tugraz.at/~hadley/ss2/landau/first_order.php (accessed
14 August 2019).
Guo, Y., Luo, H., Chen, K. et al. (2002). Effect of composition and poling field on the
properties and ferroelectric phase-stability of Pb(Mg1/3 Nb2/3 )O3 –PbTiO3
crystals. Journal of Applied Physics 92 (10): 6134–6138.
Hayashi, H., Nakamura, T., and Ebina, T. (2013). In-situ Raman spectroscopy of
BaTiO3 particles for tetragonal–cubic transformation. Journal of Physics and
Chemistry of Solids 74 (7): 957–962. Available at: http://www.sciencedirect.com/
science/article/pii/S0022369713000668.
44 2 Introduction to Ferroelectrics
Pan, W., Zhang, Q., Bhalla, A., and Cross, L.E. (1989a). Field-forced
antiferroelectric-to-ferroelectric switching in modified lead zirconate titanate
stannate ceramics. Journal of the American Ceramic Society 72 (4): 571–578.
Available at: https://doi.org/10.1111/j.1151-2916.1989.tb06177.x.
Pan, W.Y., Dam, C.Q., Zhang, Q.M., and Cross, L.E. (1989b). Large displacement
transducers based on electric field forced phase transitions in the tetragonal
(Pb0.97 La0.02 ) (Ti,Zr,Sn)O3 family of ceramics. Journal of Applied Physics 66 (12):
6014–6023. Available at: https://doi.org/10.1063/1.343578.
Qu, W. and Tan, X. (2006). Texture control and ferroelectric properties of
Pb(Nb,Zr,Sn,Ti)O3 thin films prepared by chemical solution method. Thin Solid
Films 496 (2): 383–388. Available at: http://www.sciencedirect.com/science/
article/pii/S0040609005017608.
Ren, X. (2004). Large electric-field-induced strain in ferroelectric crystals by
point-defect-mediated reversible domain switching. Nature Materials 3 (2):
91–94.
Sawaguchi, E., Maniwa, H., and Hoshino, S. (1951). Antiferroelectric structure of
lead zirconate. Physical Review 83 (5): 1078.
Scott, J.F. (2011). Electrocaloric materials. Annual Review of Materials Research 41
(1): 229–240. Available at: https://doi.org/10.1146/annurev-matsci-062910-
100341.
Shirane, G., Sawaguchi, E., and Takagi, Y. (1951). Dielectric properties of lead
zirconate. Physical Review 84 (3): 476–481.
Shirane, G., Axe, J.D., Harada, J., and Remeika, J.P. (1970). Soft ferroelectric modes
in lead titanate. Physical Review B 2 (1): 155–159.
Stock, C., Birgeneau, R.J., Wakimoto, S. et al. (2004). Universal static and dynamic
properties of the structural transition in Pb(Zn1/3 Nb2/3 )O3 . Physical Review B 69
(9): 094104.
Takezoe, H., Gorecka, E., and Čepič, M. (2010). Antiferroelectric liquid crystals:
interplay of simplicity and complexity. Reviews of Modern Physics 82 (1): 897–937.
Tan, X., Ma, C., Frederick, J. et al. (2011). The antiferroelectric ↔ ferroelectric phase
transition in lead-containing and lead-free perovskite ceramics. Journal of the
American Ceramic Society 94 (12): 4091–4107.
Tuttle, B.A. and Payne, D.A. (1981). The effects of microstructure on the
electrocaloric properties of Pb(Zr,Sn,Ti)O3 ceramics. Ferroelectrics 37 (1):
603–606. Available at: https://doi.org/10.1080/00150198108223496.
Uchino, K. (1985). Digital displacement transducer using antiferroelectrics.
Japanese Journal of Applied Physics 24: 460–462.
Uchino, K. (2009). Ferroelectric Devices, 2e. CRC Press.
Vakhrushev, S., Nabereznov, A., Sinha, S.K. et al. (1996). Synchrotron X-ray
scattering study of lead magnoniobate relaxor ferroelectric crystals. Journal of
Physics and Chemistry of Solids 57 (10): 1517–1523. Available at: http://www
.sciencedirect.com/science/article/pii/0022369796000224.
Vanderbilt, D. and Cohen, M.H. (2001). Monoclinic and triclinic phases in
higher-order Devonshire theory. Physical Review B 63 (9): 094108.
Vendik, O.G., Hollmann, E.K., Kozyrev, A.B., and Prudan, A.M. (1999). Ferroelectric
tuning of planar and bulk microwave devices. Journal of Superconductivity 12 (2):
325–338. Available at: https://doi.org/10.1023/A:1007797131173.
46 2 Introduction to Ferroelectrics
Wang, C. (2010). Theories and methods of first order ferroelectric phase transitions.
In: Ferroelectrics (ed. I. Coondoo). InTech. Available at: https://www.intechopen
.com/books/ferroelectrics/theories-and-methods-of-first-order-ferroelectric-
phase-transitions.
Xu, G., Shirane, G., Copley, J.R.D., and Gehring, P.M. (2004). Neutron elastic diffuse
scattering study of Pb(Mg1/3 Nb2/3 )O3 . Physical Review B 69: 064112.
Ye, Z.G., Noheda, B., and Dong, M. (2001). Monoclinic phase in the relaxor-based
piezoelectric/ferroelectric Pb(Mg1/3 Nb2/3 )O3 –PbTiO3 system. Physical Review B
64 (18): 184114.
Yoshida, M., Mori, S., and Yamamoto, N. (1998). Transmission electron microscope
observation of relaxor ferroelectric Pb(Mg1/3 Nb2/3 )O3 . Journal of the Korean
Physical Society 32: S993–S995.
Zhang, Q., Tan, G., Yu, R. et al. (2013). Direct observation of multiferroic vortex
domains in YMnO3 . Scientific Reports 3: 2741. Available at: http://dx.doi.org/10
.1038/srep02741.
Zhang, N., Yokota, H., Glazer, A.M. et al. (2014). The missing boundary in the phase
diagram of PbZr1−x Tix O3 . Nature Communications 5 (May): 5231.
Zhong, W. and Vanderbilt, D. (1996). Effect of quantum fluctuations on structural
phase transitions in SrTiO3 and BaTiO3 . Physical Review B 53 (9): 5047–5050.
47
Some smart systems are based on ferroelectric materials; among their many
applications are actually the implementation of their piezoelectric properties
since most excellent piezoelectric materials are also ferroelectrics. The device
applications of piezoelectric property in smart systems will be introduced
in Chapter 7. In this chapter, device applications based on the change of
ferroelectric polarization will be introduced.
Ferroelectric film
– – – – – – – – – –
Bit line
Word line
Word line
Bit line
N N
P
Plate line
(a)
1T
VR
1V FBB/
RBB
ID ≈ 100
μ A
Low-VT state
(b)
Figure 3.1 (a) FeRAM device and structural diagram (1T1C) made by Fujitsu, where the
ferroelectric layer’s polarization switching induces large difference in capacitance and
therefore the source–drain current representing “0” and “1” states. Source: Redrawn from
Fujitsu (n.d.). (b) FeRAM (or FeFET) from the Ferroelectric Memory Company and structural
diagram showing 1T structure. Source: Adapted from Ferroelectric Memory Company (2018)
and Arimoto and Ishiwara (2004).
inside the polarized layer by the generation of surface charge. This field’s direction
is opposite to the field that induces polarization and tends to depole the ferroelec-
tric layer; therefore, it is called depolarization field as illustrated in Figure 3.2.
Figure 3.3 shows 1T FeRAM diagram, where the gate dielectric layer is fer-
roelectric with polarization up and down representing the “0” and “1” states.
Figure 3.3a illustrates that the surface charge induced by ferroelectric polariza-
tion results in the shift of ON/OFF threshold voltage of field effect transistor
(FET) exhibiting a memory window.
The earlier studies of FeRAM were focused on PZT, since it has much larger
remnant polarization compared with other ferroelectric materials such as
BaTiO3 and PbTiO3 . However, PZT film faces serious fatigue problem, i.e. after
108 times of polarization switching, its ferroelectricity decays significantly. This
is not acceptable for non-volatile memory application (Al-Shareef et al. 1994).
This problem has hindered scientists quite a long time but attracted a lot of
interest of study. Later on, an alternative ferroelectric material SrBi2 Ta2 O9 (SBT)
has been found to be fatigue-free up to 1011 cycles of polarization switching.
3.1 Ferroelectric Random-Access Memory 49
P0 Poling field
Ed
Gate
– – – –
Spontaneous polarization
of ferroelectric layer Memory
–––––––––
window
Source Drain
– – – – – – – – – – IDS
N N –––––––––
P
– – – –
VG
(a) (b)
Figure 3.3 (a) Diagram of the 1T FeRAM based on a field-effect transistor with ferroelectric
gate dielectric. (b) The ferroelectric polarization-induced interfacial charge alters the on–off
voltage of the FET resulting “0” and “1” states.
SBT also has a relatively smaller coercive field making it able to be operated with
lower voltage.
BiFeO3 (BFO) came into the focus of many scientists in the ferroelectric field
due to the discovery of its very large remnant polarization in thin film form,
which makes it very promising for FeRAM. However, it is very difficult to grow
high-quality BFO thin films with low leakage current and low dielectric loss. This
hinders the mass production of BFO-based FeRAM (Ielmini and Waser 2016).
An emerging ferroelectric material for FeRAM application is HfO2 , which has
been found to be ferroelectric when doped with ZrO2 or other oxides. Even the
ferroelectricity of HfO2 is usually not as good as PZT and BiFeO3 , but since it has
been implemented in CMOS technology as gate insulator, FeRAM with HfO2
gate oxide should be very compatible to current COMS process and therefore is
very attractive to semiconductor industry. However, the origin of ferroelectricity
is still not very clear, even it has been practically demonstrated to be ferroelectric.
Details will be introduced in Chapter 5.
50 3 Device Applications of Ferroelectrics
δ1 Φ– Φ+
EF EF
δ2
M1 FE M2 M1 FE M2
P P
– + + –
Figure 3.4 Energy band diagram of a ferroelectric tunnel junction. Polarizations in two
opposite directions cause two average barrier heights resulting in two resistance states. A
rectangular barrier, denoted by dashed lines is assumed when ferroelectric is nonpolarized.
Source: Garcia and Bibes (2014). Reprinted with permission of Springer Nature.
ER
5.0 1.0
TE 105 TE TE
102 FNT 4.0 0.5
j (A/cm2)
d (nm)
10–1
FNT
FNT
FNT 3.0 0.0
10–4 DT
DT
10–7 2.0 –0.5
DT TE
.5
.0
.5
0
5
1.0
1.5
0.
0.
V (V)
–1
–1
–0
(a) (b) (c) V (V)
Figure 3.5 (a) Illustration and (b) voltage dependence of the current density contributions for
3.2 nm-thick BaTiO3 of three possible transport mechanisms through ultrathin ferroelectrics:
direct tunneling (DT), Fowler–Nordheim tunneling (FNT), and thermionic emission (TE). Solid
and dashed lines stand for two polarization orientations. (c) Corresponding electroresistance
(ER) as a function of voltage and ferroelectric thickness d. The dashed lines illustrate the
boundaries where each DT, FNT, and TI mechanism dominates. Source: Pantel and Alexe
(2010). Reprinted with permission of American Physical Society.
IDT = SC [( )1 ( )1 ]
𝛼2 𝜑2 − eV2 − 𝜑1 + eV2
2 2
{ [( )1 ( ) 1 ]}
3eV eV 2 eV 2
× sinh 𝛼 𝜑2 − − 𝜑1 + (3.4)
4 2 2
with
32𝜋em∗
C=− (3.5)
9h3
and
1
8𝜋d(2m∗ ) 2
𝛼= (3.6)
3h(𝜑1 + eV − 𝜑2 )
where I DT is the direct tunneling current given by Gruverman et al. (2009), S the
area of the interface, d the film thickness, h the Planck constant, m* the effective
carrier mass, and Φ1 and Φ2 are potential steps at the two interfaces.
Similar to DT, FNT shares the same physical phenomenon but in different
voltage regime. It is the process whereby electrons tunnel through a barrier in
the presence of a high electric field. Across a triangular-shaped potential barrier,
FNT is present. The tunnel probability can be derived from time-independent
Schrödinger equation, and the calculation of current yields the following
3.2 Ferroelectric Tunneling Non-volatile Memory 53
relationship between the current density J FN and the electric field in the barrier E:
( √ 3 )
2m ∗ (e𝜑 ) 2
4 B
JFN = CFN E2 exp − (3.7)
3 eℏ E
where 𝜑B is the barrier height at the conductor/barrier interface.
TE describes the electron flow as the charge carriers that overcome the poten-
tial barrier by thermal energy. Usually, the I–V curves can be fitted by TE model.
The forward bias currents are described by the theory of TE as the following
equation:
( e𝜑 ) [ ( ) ]
eV
Iforward = SA∗ T 2 exp − B exp −1 (3.8)
kT nkB T
where S is the junction area, A* the Richardson constant, T the absolute temper-
ature, e the electron charge, k the Boltzmann constant, n the ideal factor, and 𝜑B
the Schottky barrier height. From the linear relationship in TE modeling, we can
identify the barrier height in OFF state using Eq. (3.8). The corresponding deple-
tion region width, where electrons are depleted in interface, can be estimated by
Eq. (3.9):
[ ]1
2𝜖s 2
WD = (Vbi − V ) (3.9)
eND
where W D is the depletion region width, 𝜖 s the dielectric constant of semicon-
ductor, N D the donor concentration, V bi = 𝜑B + 𝜑n , and 𝜑n = EF − Ec .
In Figure 3.5b, the current density as a function of voltage across 3.2-nm thick
ferroelectric thin film is shown. For the two opposite polarizations (P > 0 and
P < 0 with solid and dashed lines, respectively), the current density behaves
differently in three mechanisms. DT is prominent in low voltage regime, while
FNT is in high voltage regime. With a small range of voltage, DT gives a parallel
current branch for both polarization directions showing the independence of
applied voltage on electroresistance. On the other hand, FNT is barely observed
in low voltage. Because of insufficient band tilting, there is virtually no current
recorded in low voltage regime. In the case of TE, electroresistance is obvious
as its value is given by the change in barrier height on polarization reversal.
However, the total current density decreases compared with the other two.
Therefore, it is usually suitable for thicker films that tunneling cannot happen
(see Figure 3.5c).
Among these three mechanisms, we can understand the possibility in elec-
tronic transportation across ferroelectric thin film. Certainly, parameters related
to those mechanisms, such as effective mass of oxide film in direct tunneling and
FNT as well as temperature in TE, can shift between mechanisms. I–V mea-
surement is the first step to see a ferroelectric junction’s memory characteristics.
However, I–V curves alone are not sufficient to clarify the underlying switching
mechanism that should be attributed to polarization switching.
By performing sequential ferroelectric poling, piezoresponse force microscopy
(PFM) imaging, and conductive-atomic force microscope (C-AFM) imaging
of the poled domains, the resistance difference measured with the C-AFM or
54 3 Device Applications of Ferroelectrics
other electrical methods between positively and negatively poled regions would
therefore correspond to the ferroelectric-induced ER effect (Gruverman et al.
2009). Garcia et al. (2009) investigated the local tunneling electroresistance
based on this method and demonstrated resistive readout of the polarization
state through its influence on the tunnel current using PFM and C-AFM at room
temperature even the film is a 1 nm-highly strained BaTiO3 film. The resulting
electroresistance effect scales exponentially with ferroelectric film thickness
and it can reach 75 000% at 3 nm thick. In fact, only by tunneling at such thin
thickness is able to read the polarization state without destroying it, otherwise
undesirable leakage current will dominate. Apart from using conductive tips as
top electrodes, Chanthbouala et al. (2012b) reported tunnel electroresistance on
Au/Co/BaTiO3 (2 nm)/LSMO(30 nm)/NdGaO3 heterostructure with resistance
switching occurring at the coercive voltage of ferroelectric switching. They
show an alternative way to ensure that the electroresistance is originated from
ferroelectric response (Figure 3.6).
heterostructure showing
0 ferroelectric switching versus
resistive switching. (c)
–45 Resistance versus writing
voltage measured in
–90 remanence after applying
(a) successive voltage pulses. The
open and filled circles
represent two different scans
Amplitude (a.u.)
to show reproducibility.
10
Source: Adapted from
Chanthbouala et al. (2012b).
1
(b)
107
Resistance (Ω)
106
105
(c)
–4 –3 –2 –1 0 1 2 3 4
Vwrite (V)
3.2 Ferroelectric Tunneling Non-volatile Memory 55
M F S 107
P
wr
w
106 OFF
EF
Resistance (Ω)
EC
EV 105
ON OFF
M F S 104
ON
r w
103
EF P
EC w
EV
–5 –4 –3 –2 –1 0 1 2 3 4 5
(a) (b) Voltage (V)
1.0 V
1.5 V
2.0 V
1000 2.5 V
OFF
Resistance (kΩ)
100
ON
10
–4 –3 –2 –1 0 1 2 3 4
(a) Voltage (V)
2.5 V 0V 0V
(b)
Figure 3.8 (a) Tunneling resistance with varying maximum (positive or negative) writing
voltages and (b) configuration extracted from the PFM phase images showing the relative
fraction of ferroelectric domains pointing downward with increased voltage. Source:
Reprinted with permission from Yau et al. (2015).
3.2 Ferroelectric Tunneling Non-volatile Memory 57
107
Pt 106
BTO
105
NSTO
ON/OFF ratio
104
In-plane
strain ε
103
4.30%
2.18% 102
101
0.18%
c 100
Ba 2+ –4.0 –3.0 –2.0 –1.0 0.0
Ti 4+
a O 2– Strain ε (%)
(a) (b)
Figure 3.9 (a) Illustration of in-plane strain effect and (b) the corresponding ON/OFF ratio of
Pt/BaTiO3 /Nb:SrTiO3 metal/ferroelectric/semiconductor heterostructured ferroelectric tunnel
junction. Source: Adapted from Yau et al. (2017).
58 3 Device Applications of Ferroelectrics
Thermometer
Electrometer
Thermocouple
Sample
Lock-in amplifier
Signal-out Reference Signal-in
Pyroelectric coefficient
Material (𝛍C/(m2 K)) References
– – – – – – – – – – – – – – – – – –
+ + + + + + + + + + + + + + + + + +
+ + + – + – + + + + – +
+ + +
P – – – P – –– P – – –
– – – – – – – – – – – – – – – – – –
+ + + + + + + + + + + + + + + + + +
IP IP
dT dT dT
0 0 0
dt dt dt
(a) (b) (c)
Figure 3.11 (a) Diagram of pyroelectric infrared sensor, (b) decreased polarization and change
of surface charges when thermal radiation is on, and (c) increased polarization and change of
surface charges when thermal radiation is off. The ellipsoids are only guidance for electric
polarization change.
+5 V
Amplifier
Pyroelectric
elements R1 R2
(a) (b)
Figure 3.12 (a) Basic structure and (b) circuit of a pyroelectric detector.
element to fabricate an infrared (IR) sensor, many factors such as the operat-
ing temperature, frequency, and pyroelectric response should be considered. It
is possible to formulate a group of FOMs to assess the performance of IR device.
The current FOM, F i , which describes the influence of the pyroelectric material
on the current responsibility of a pyroelectric sensor, is defined as
p
Fi = (3.14)
𝜌Cp
where p is the pyroelectric coefficient, 𝜌 is the density of the sample, and C p is
the heat capacity.
The voltage FOM, F v , which evaluates the influence of the pyroelectric-material
on the voltage responsibility of a pyroelectric sensor, is expressed as
p
Fv = (3.15)
𝜌Cp 𝜖0 𝜖r
where 𝜖 0 and 𝜖 r are the permittivity of free space and relative permittivity of the
sample, respectively.
The detectivity FOM, F d assesses the influence of the pyroelectric material on
the detection sensitivity of a pyroelectric sensor and is given by
p
Fd = √ (3.16)
𝜌Cp 𝜖0 𝜖r tan 𝛿
where tan 𝛿 is the dielectric loss tangent.
This equation tells us that to make a pyroelectric sensor with high performance,
the pyroelectric coefficient should be as large as possible; other parameters such
as heat capacity, dielectric permittivity and dielectric loss, and specific weight
should be the smaller the better.
PMN-PT single crystal has superior pyroelectric performances to make it
a promising candidate for high-performance uncooled infrared detectors and
thermal imaging. From Table 3.1, the pyroelectric coefficient of PMN-PT single
crystal is proven to be larger than some classical ferroelectric materials such as
PZT ceramic. However, the measured pyroelectric coefficients of PMN-PT are
62 3 Device Applications of Ferroelectrics
far away from another Mn-doped PMN-PT single crystal. Hence, the pyroelec-
tric coefficient is varied from composition and crystal orientation. (Students:
Think how to choose the crystal orientation for largest pyroelectric response.)
Table 3.2 summarizes PMN-PT materials’ group of FOMs with comparison of
the other ferroelectric materials. We can see that PMN-PT relaxor ferroelectric
crystal has superior FOM compared with the ever best pyroelectric LiTaO3 (Li
et al. 2014; Tang and Luo 2009; Tang et al. 2005, 2006; Liu et al. 2009; Kulwicki
et al. 1992).
Another method to characterize the pyroelectric effect is to use infrared laser
with a chopper and lock-in amplifier. As illustrated in Figure 3.13, an infrared
sensor based on pyroelectric effect can be characterized with infrared radiation
from an incident of infrared laser light, and current or charges can be measured
Pyroelectric materials
Absorption layer
IR laser
A
Chopper
Lock-in
amplifier
Signal generator
Lock-in
amplifier
MoS2
Au e– e– e– e– e– e– Au
MoS2
PMN-PT
S D – – – – – –
C/Au PMN-PT C/Au
MoS2
ITO Au e– e– e– Au
PMN-PT
IR – – –
C/Au
(a) (b) IR
0.018
6 mW/mm2 IR off VD = 0.5 V
0.016
Current (μA)
0.014
0.012
0.010
IR on
0.008
0 200 400 600 800
(c) Time (s)
Figure 3.14 (a) A schematic view of the ferroelectric gate FET, (b) the working mechanism of
the ferroelectric field-effect transistor modulated with IR illumination, and (c) the
time-resolved photocurrent in response to IR on/off at an irradiance of 6 mW/mm2 with
1064 nm laser. Source: Reprinted with permission from Fang et al. (2015). Copyright 2015, The
Optical Society.
by lock-in technique. In the setup, the alternative laser incident can be realized
by a laser chopper or by a laser beam excited by sinusoidal or square wave of
electrical input that is also used as the reference input to the lock-in amplifier.
Figure 3.14 shows our result of an example to integrate pyroelectric effect
with a semiconducting 2D material to form an infrared sensor. This device uses
pyroelectric-induced field effect to modulate the surface conductivity of 2D
MoS2 to enable it to detect infrared radiation, or otherwise it is only sensitive to
visible or UV light.
BST thin films have been employed in many applications, for example, the BST
tunable capacitor used for matching networks in antenna and amplifier. By chang-
ing its capacitance as bias voltage changes, it can tune antennas, improve the
effectiveness of the phone’s antenna, and help the phone to avoid dead spots in
order to improve performance especially in regions with low signal. The elec-
tronically steerable system saves space and weight in satellite communication and
radar system. Not limited to communication technology, BST thin films also have
the potential in integration with semiconductor microelectronic circuits, includ-
ing microwave tunable switches and resonators (Figure 3.15) (Terrence 2013).
Although BST is currently competing with several materials used in the
market, such as semiconductor GaAs and micro-electrical–mechanical sys-
tems (MEMS), the outstanding properties of BST still shows its value in the
development. Table 3.3 shows the comparison of the properties of GaAs, MEMS,
BST Gap
Patch
Substrate
Ground plane
0
Reflection magnitude (dB)
–5
–10
BL- Corner
–15
BL- Emin
BL- Middle
–20 No BL
–25
9 9.5 10 10.5 11 11.5 12
Frequency (GHz)
(a) (b)
Figure 3.15 (a) Antenna devices with tunable BST layer acting as a distributed capacitance
and corresponding reflection (Karnati et al. 2015). (b) The wafer-level antenna devices with BST
thin films. Source: Adapted from Terrence (2013).
Table 3.3 Comparison of the properties of GaAs, MEMS, and BST thin film.
and BST. We can see that GaAs and BST show the highest tunability and lowest
dielectric loss even though MEMS is at a similar level of development maturity
as BST. However, the linear decrease of the quality factor with frequency has
limited the applications of GaAs in tunable filters for f > 1 GHz. Additionally,
they are more expensive and consume more power than ferroelectric thin films
in fabrication and operation.
Light
Top electrode
Ferroelectric
– – – – –
P E thin film
+ + + + +
Bottom electrode
Hole
Electron
Figure 3.16 Schematic illustration of the physical mechanism of the photovoltaic effect in a
ferroelectric.
66 3 Device Applications of Ferroelectrics
ITO
Al
Al
Polymer: PCBM
P(VDF-TrFE)
Figure 3.17 The structure of polymer photovoltaic devices with FE interfacial layers and a
schematic diagram of electric field induced by the polarized FE layer and the field-assisted
charge extraction. The right-hand panel illustrates the electric-field distribution and electron
conduction through the P(VDF-TrFE) on the Al side. Source: Yuan et al. (2011). Reprinted with
permission of Springer Nature.
200
Dark
Light
150 A Polarization up
ITO Polarization down
Current density (nA/cm2)
100 v
BNBT
50 ITO
–50
–100
–150
–200
–100 –75 –50 –25 0 25 50 75 100
Voltage (V)
Figure 3.18 Current density–voltage (J–V) curve of BNBT:Nb ceramics under dark and UV
illumination of 365 nm. Inset is the configuration for photovoltaic measurement. Source: Chen
et al. (2014). Reprinted with permission of Elsevier.
Figure 3.19 is the diagram proposed by Huang to illustrate the working princi-
ple of the photovoltaic effect compared with a conventional p–n junction solar
cell (Huang 2010). We can see that the band bending at the domain wall is favor-
able for charge traps and therefore enhance the charge separation. However, we
have to face the truth: a good ferroelectric requires very good insulating property,
3.6 Electrocaloric Effect 67
hv VB hv
VB
P P P
n-Type Depletion layer p-Type
(a) (b)
Figure 3.19 Principles of operation for a conventional solar cell and the ferroelectric
alternative. (a) Band diagram of a conventional semiconductor p–n junction, and (b) in the
ferroelectric domain, band bending occurs across the domain walls (DWs). The directions of
electric polarization (P) are shown by arrows. Source: Adapted from Huang (2010).
E=0
Heat release Heat absorb
Heat irradiation
– – – – – –
E>0
Figure 3.20 (a) An illustration of the basic principle of the electrocaloric effect in a polar
material and (b) a multilayered electrocaloric device where the electrocaloric films are
ferroelectric. Source: Adapted from Epstein and Malloy (2010).
References
Al-Shareef, H.N., Bellur, K.R., Auciello, O., and Kingon, A.I. (1994). Fatigue and
retention of Pb(Zr0.53 Ti0.47 )O3 thin film capacitors with Pt and RuO2 electrodes.
Integrated Ferroelectrics 5 (3): 185–196. Available at: https://doi.org/10.1080/
10584589408017011.
Arimoto, Y. and Ishiwara, H. (2004). Current status of ferroelectric random-access
memory. MRS Bulletin 29 (11): 823–828.
Bao, P., Jackson, T.J., Wang, X., and Lancaster, M.J. (2008). Barium strontium
titanate thin film varactors for room-temperature microwave device applications.
Journal of Physics D: Applied Physics 41 (063001): 21.
Buser, R.G., Tompsett, M.F., Kruse, P.W. et al. (1997). Ch.4 Hybrid
Phyroelectric-Ferroelectric Bolometer Arrays. In: Uncooled Infrared Imaging
Arrays and Systems, 1e (ed. C.M. Hanson), 154–168. Academic Press.
Cao, D., Zheng, F., Shen, M. et al. (2012). High-efficiency ferroelectric-film solar
cells with an n-type Cu2 O cathode buffer layer. Nano Letters 12 (6): 2803–2809.
Chanthbouala, A., Garcia, V., Barthélémy, A. et al. (2012a). A ferroelectric
memristor. Nature Materials 11: 860–864. Available at: http://www.ncbi.nlm.nih
.gov/pubmed/22983431.
Chanthbouala, A., Crassous, A., Garcia, V. et al. (2012b). Solid-state memories
based on ferroelectric tunnel junctions. Nature Nanotechnology 7 (2): 101–104.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/22138863.
References 69
Chen, F.S. (1969). Optically induced change of refractive indices in LiNbO3 and
LiTaO3 . Journal of Applied Physics 40 (8): 3389–3396. Available at: https://doi
.org/10.1063/1.1658195.
Chen, L., Luo, B.C., Wang, D.Y. et al. (2014). Enhancement of photovoltaic
properties with Nb modified (Bi,Na)TiO3 –BaTiO3 ferroelectric ceramics. Journal
of Alloys and Compounds 587: 339–343. Available at: http://www.sciencedirect
.com/science/article/pii/S0925838813026339.
Choi, T., Lee, S., Cheong, S.W. et al. (2009). Switchable ferroelectric diode and
photovoltaic effect in BiFeO3 . Science 324 (5923): 63–66. Available at: http://
science.sciencemag.org/content/324/5923/63.
Epstein, R.I. and Malloy, K.J. (2010). Electrocaloric refrigerator and multilayer
pyroelectric energy generator, Pub. No.: US 2010/0037624 A1. pp. 1–10. Available
at: https://patents.google.com/patent/US20100037624.
Fang, H., Yan, Q., Dai, J.Y. et al. (2015). Infrared light gated MoS2 field effect
transistor. Optics Express 23 (25): 31908–31914. Available at: http://www
.opticsexpress.org/abstract.cfm?URI=oe-23-25-31908.
Ferroelectric Memory Company (2018). One-transistor FeFET memory. Available
at: https://ferroelectric-memory.com/technology/one-transistor-fefet-memory/
(accessed 19 August 2019).
Fujitsu (n.d.). FRAM structure. Available at: http://www.fujitsu.com/global/
products/devices/semiconductor/memory/fram/overview/structure/ (Accessed
14 March 2018).
Garcia, V. and Bibes, M. (2014). Ferroelectric tunnel junctions for information
storage and processing. Nature Communications 5: 4289. Available at: http://
www.ncbi.nlm.nih.gov/pubmed/25056141.
Garcia, V., Fusil, S., Bibes, M. et al. (2009). Giant tunnel electroresistance for
non-destructive readout of ferroelectric states. Nature 460 (7251): 81–84.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/19483675.
Gruverman, A., Wu, D., Lu, H. et al. (2009). Tunneling electroresistance effect in
ferroelectric tunnel junctions at the nanoscale. Nano Letters 9 (10): 3539–3543.
Gu, H., Qian, X.S., Ye, H.J., and Zhang, Q.M. (2014). An electrocaloric refrigerator
without external regenerator. Applied Physics Letters 105 (16): 162905. Available
at: https://doi.org/10.1063/1.4898812.
Huang, H. (2010). Solar energy: ferroelectric photovoltaics. Nature Photonics 4:
134–135.
Ielmini, D. and Waser, R. (eds.) (2016). Resistive Switching: From Fundamentals of
Nanoionic Redox Processes to Memristive Device Applications. Weinheim,
Germany: Wiley-VCH Verlag GmbH & Co.
Karnati, K.K., Shen, Y., Trampler, M.E. et al. (2015). A BST-integrated capacitively
loaded patch for Ka- and X-band beamsteerable reflectarray antennas in satellite
communications. IEEE Transactions on Antennas and Propagation 63 (4):
1324–1333.
Kulwicki, B.M., Amin, A., Beratan, H.R., and Hanson, C.M. (1992). Pyroelectric
imaging. In: ISAF’92: Proceedings of the Eighth IEEE International Symposium on
Applications of Ferroelectrics, 1. IEEE https://doi.org/10.1109/ISAF.1992.300607.
Li, L., Zhao, X., Luo, H. et al. (2014). Scale effects of low-dimensional relaxor
ferroelectric single crystals and their application in novel pyroelectric infrared
70 3 Device Applications of Ferroelectrics
Yau, H.M., Yan, Z.B., Dai, J.Y. et al. (2015). Low-field switching four-state nonvolatile
memory based on multiferroic tunnel junctions. Scientific Reports 5: 12826.
Available at: http://www.nature.com/doifinder/10.1038/srep12826.
Yau, H.M., Xi, Z., Dai, J.Y. et al. (2017). Dynamic strain-induced giant
electroresistance and erasing effect in ultrathin ferroelectric tunnel-junction
memory. Physical Review B 95: 214304.
Yuan, Y., Reece, T.J., Huang, J. et al. (2011). Efficiency enhancement in organic solar
cells with ferroelectric polymers. Nature Materials 10: 296–302.
Zhuravlev, M., Sabirianov, R., Jaswal, S., and Tsymbal, E. (2005). Giant
electroresistance in ferroelectric tunnel junctions. Physical Review Letters 94 (24):
246802. Available at: http://link.aps.org/doi/10.1103/PhysRevLett.94.246802.
73
Ferroelectric Characterizations
In recent years, many new emerging materials are being claimed to be ferroelec-
tric, such as two-dimensional (2D) materials MoS2 , SnSe, and topological insula-
tor material CsPbI3 (Liu et al. 2016). Even the solar cell material hybrid perovskite
CH3 NH3 PbO3 has been claimed to be ferroelectric, which is also attributed to be
one of the main reasons to facilitate the photovoltaic efficiency, but the result is
still under debate. To clarify that a material is ferroelectric needs critical exper-
imental evidence such as P–E loop and switchable electric polarization. In this
chapter, basic techniques and recent advances of ferroelectric characterization
methods are introduced.
Sample R
Cx Rx
Cref
Figure 4.1 A modified Sawyer–Tower circuit for ferroelectric P–E loop measurement.
where V ref is the voltage for the reference capacitor and V x is the voltage for the
sample. Given the dimension of the sample area S, thickness d, and dielectric
constant 𝜖 r , under electric field E, the sample’s ferroelectric polarization can be
found by
Cx 𝜖𝜖 SV 𝜖𝜖S S
Vref = V = r 0 x = r 0 E= P (4.3)
Cref x Cref d Cref Cref
Then, we have
Vref Cref
P= (4.4)
S
During the hysteresis loop measurement, an ac signal generated by a function
generator and amplified by a voltage amplifier is applied to the sample placed
in a silicone oil bath. The input and output signals of Sawyer–Tower circuit are
observed by a digital oscilloscope, and the data of hysteresis loop at different tem-
peratures are recorded in a computer.
As an example, Figure 4.2 shows our experimental result of P–E loop measure-
ment, from which we can determine that the remnant polarization Pr = 0.25 C/m2
or 25 μC/cm2 and saturation polarization Ps = 0.29 C/m2 or 29 μC/cm2 . The coer-
cive field Ec = 1.80 × 105 V/m.
To obtain a P–E loop of a typical ferroelectric material is not difficult; however,
for many ferroelectric materials with large leakage current (especially thin films),
it is not an easy task. Figure 4.3a sketch an ideal linear response of a capacitor
in the P–E loop measurement, and Figure 4.3b shows an ideal resistor response.
For a leaky ferroelectric thin film, the measured hysteresis loops sometimes only
show ellipsoid feature without saturation as illustrated in Figure 4.3c. This kind
of loop cannot be called P–E loop since it only reflects a leaking dielectric char-
acteristic, and therefore, this kind of result cannot support the conclusion of
ferroelectric claim of the corresponding materials (Scott 2008). A successful P–E
loop should present at least saturation in the loop as shown in Figure 4.3d.
4.1 P–E Loop Measurement 75
0.2
PR = 0.25 C/m2
Polarization (C/m2)
0.1
0.0
Ec = 1.80 × 105 V/m
–0.1
–0.2
–0.3
–1.5 × 106 –1.0 × 106 –5.0 × 105 0.0 5.0 × 105 1.0 × 106 1.5 × 106
Electric field (V/m)
Figure 4.2 Typical P–E loop from a PMN-PT relaxor ferroelectric single crystal.
P P
E E
(a) (b)
P P
E E
(c) (d)
Figure 4.3 (a) An ideal linear response of capacitor, (b) an ideal resistor response, (c) an
unsuccessful “P–E loop” without a saturation, which is a typical dielectric loss induced loop,
and (d) a typical P–E loop showing ferroelectricity.
76 4 Ferroelectric Characterizations
ϵr (T)
25 000
x=0
x = 0.34
20 000 x = 0.27
x = 0.47 x = 0.65
x = 0.87
15 000
10 000 x = 1.0
5000
0
0 100 200 300 400
(a) Temperature (K)
PMN-PT
2
104ϵ
Frequency
increases
0
–100 –50 0 50 100
Figure 4.4 Dielectric–temperature curve of (a) (Bax Sr1−x )TiO3 where x varies from 0 to 1, i.e.
from SrTiO3 to BaTiO3 and (b) PMN-xPT. Source: Adapted from Vendik et al. (1999) and Cheng
et al. (1998).
4.3 Piezoresponse Force Microscopy (PFM) 77
curves of (Bax Sr1−x )TiO3 (BST) show different Curie temperatures for different
components of x. The 𝜖 r –T curves at different frequencies also present the
frequency dispersion if a ferroelectric is relaxor type such as PMN-PT as shown
in Figure 4.4b.
Photodiode detector
Laser source
Low-pass filter
Feedback circuit
Piezo-material
Scanner
Lock-in amplifier
Reference signal
Amplitude, A Phase, φ
[111]
(a) (b)
Not only the domain pattern, when an area is scanned with a conductive APM
tip being added with a dc electric field, but also the polarization of the scanned
area can be switched. Therefore, the PFM technique can also demonstrate polar-
ization switching, which is the key criterion for a ferroelectric. Furthermore, a
so-called butterfly loop (see Figure 4.7) can be achieved by PFM technique when
the piezoresponse amplitude is measured as a function of dc electric field added
on the sample. The butterfly loop illustrates that the increased dc field results
increased polarization leading to larger piezoresponse. It is worthy to note that
the nonzero piezoresponse at zero field is due to the remnant polarization, and
the corresponding phase angle change from 0∘ to 180∘ is due to the polariza-
tion switching. Therefore, this butterfly loop and the corresponding 180∘ phase
change can be used as criteria for clarification of ferroelectricity.
To demonstrate ferroelectric switching is essential to claim ferroelectricity. In
PFM imaging, when a dc voltage is applied on the conductive AFM tip while scan-
ning an area, the subsequent PFM image can reveal the switched polarization as
illustrated in Figure 4.8, where (a) is the amplitude image and (b) the phase image.
4.3 Piezoresponse Force Microscopy (PFM) 79
Phase (°)
100
50
0
Amplitude (a.u.)
–6 –4 –2 0 2 4 6
Voltage (V)
a.u. °
1 180
1 μm 1 μm
0 0
(a) (b)
Figure 4.8 Our PFM results from epitaxial BaTiO3 thin films. (a) Amplitude and (b) phase
images recorded after writing an area of 3.0 × 3.0 μm2 with −4.5 V and then the central square
with +4.5 V using a biased conductive tip. Source: Yau et al. (2017). Reproduced with
permission of American Physical Society.
80 4 Ferroelectric Characterizations
For the switched areas with polarizations facing up and down, the brightness of
amplitude image represents the strength of piezoresponse, while the phase image
represents the polarization direction.
Torsion Deflection
Out-of-plane
E P E P P
A B
In-plane
a-domain c-domain
C D
X3 X1
PP X2
E E P X2
X1 X3
Figure 4.9 (a) Illustration of four sectors of a photo-detector in AFM. (b) Quantitative
explanation of the contrast formation in IPP- and OPP-PFM. The electric field induced from a
conductive tip causes surface displacement due to the converse piezoelectric effect. Source:
Adapted from Felten et al. (2004).
4.3 Piezoresponse Force Microscopy (PFM) 81
IPP
IPP IPP
IPP
OPP
(d) (e)
Figure 4.10 Piezoresponse (a) OPP phase, (b) OPP amplitude, (c) IPP amplitude images of the
as-grown (001)-cut PMN-0.38PT single crystal with scan size of 5 μm, (d) the schematic
diagram of tetragonal unit cell with six polar directions, and (e) illustration of proposed
domain boundary configuration revealed from the OPP- and IPP-PFM images. Source: Adapted
from Wong et al. (2006).
polarization (pointing out of the paper) appears dark in contrast, and regions
with downward polarization appear bright (please notice that the orientation
depends on the direction of electric field).
One can see that both upward and downward domains appear dark and bright
alternatively in the OPP phase image (Figure 4.10a); while they appear bright in
OPP amplitude image (Figure 4.10b). It is interesting to notice that in the OPP
amplitude image, the dark contrast regions with the width of around 40–90 nm
at the domain boundaries, however, appear bright in the IPP amplitude image
shown in Figure 4.10c. It indicates that the domain boundaries contain large
IPP without OPP. If an IPP phase image is formed, the IPP direction can also
be mapped as bright and dark contrasts. According to a schematic diagram of a
tetragonal unit cell with six polar directions (Figure 4.10d), it can be seen that the
dark and bright stripe-like domains in the OPP phase image represent two oppo-
site OPPs, and four IPPs exist in the domain boundaries. Figure 4.10e shows the
proposed domain boundary configuration revealed from the OPP- and IPP-PFM
images.
In order to understand the nature of PFM imaging, finite element modeling
(FEM) has been carried out to illustrate the surface deformation under a finite
point electrode induced electric field. In the FEM simulation with commercial
software ANSYS, the size of the sample is set as 2 mm × 3 mm × 50 μm, which
82 4 Ferroelectric Characterizations
is same as the real sample in PFM imaging. A metal layer under the bottom of
sample as bottom electrode is fixed. Because the size of the conductive tip is very
small, the probing electric field is strongly inhomogeneous and decays rapidly
from the tip into the crystal. Therefore, we can substitute the electric field with
the same magnitude as being used in the PFM experiment into the simulation.
The displacements of two samples possessing OPP and IPPs are computed in the
ANSYS simulation.
In the simulation results (Wong 2007), the light blue and purple curves in
Figures 4.11 and 4.12 represent the applied ac electric field across the sample and
the calculated displacement in the x, y, and z directions, respectively. From the
OPP simulation result, it can be seen that along the direction of OPP, denoted
as the z-direction, there is a large displacement as shown in Figure 4.11d, while
there is no obvious displacement along the x and y directions in Figure 4.11b,c.
By contrast, in the IPP simulation result, large displacement along the x direction
can be observed as shown in Figure 4.12b; but there is no obvious displacement
along the y and z directions. Therefore, according to the movement of sample
surface, the IPP and OPP can be detected by PFM tip.
6.25
5
3.75
Out-of-plane ac 2.5
polarization field z 1.25
Value
0
–1.25
x –2.5
–3.75
y
–5
–6.25 (×10−6)
0 –2 –4 –6 –8 1
–1 –3 –5 –7 –9
(a) (b) Time
6.25 6.25
5 5
3.75 3.75
2.5 2.5
1.25 1.25
Value
Value
0 0
–1.25 –1.25
–2.5 –2.5
–3.75 –3.75
–5 –5
–6.25 (×10−6) –6.25 (×10−6)
0 –2 –4 –6 –8 1 0 –2 –4 –6 –8 1
–1 –3 –5 –7 –9 –1 –3 –5 –7 –9
(c) Time (d) Time
Figure 4.11 (a) The sample surface movement of out-of-plane polarization ANSYS simulation.
(b–d) The graph of x, y, and z displacements, respectively. The applied ac electric field is along
the out-of-plane direction. Source: Adapted from Wong (2007).
4.3 Piezoresponse Force Microscopy (PFM) 83
6.25
ac field 5
3.75
2.5
z 1.25
Value
0
–1.25
In-plane x –2.5
polarization
–3.75
y
–5
–6.25 (×10−6)
0 –2 –4 –6 –8 1
–1 –3 –5 –7 –9
(a) (b) Time
6.25 6.25
5 5
3.75 3.75
2.5 2.5
1.25 1.25
Value
Value
0 0
–1.25 –1.25
–2.5 –2.5
–3.75 –3.75
–5 –5
–6.25 (×10−6) –6.25 (×10−6)
0 –2 –4 –6 –8 1 0 –2 –4 –6 –8 1
–1 –3 –5 –7 –9 –1 –3 –5 –7 –9
(c) Time (d) Time
Figure 4.12 (a) The sample surface movement of in-plane polarization ANSYS simulation.
(b–d) The graph of x, y, and z displacements, respectively. The applied ac electric field is along
the out-of-plane direction. Source: Adapted from Wong (2007).
During PFM imaging, in order to produce an ac electric field across the sample,
an ac voltage is applied to the bottom electrode; while the conductive tip is
electrically grounded, or vice versa. Since there is an air gap between cantilever
and sample surface, capacitive force between the cantilever and sample surface
needs to be considered in PFM imaging. The capacitive force is an attractive
force between the cantilever and sample surface that alters the PFM image and
shadows out the information of domain boundary structure. Hong et al. (2002)
have investigated the capacitive force effect during PFM imaging and discovered
that, at the center position, the attractive capacitive force leads to a vertical shift
of amplitude signal in the upward and downward domains. Figure 4.13 illustrates
the electrostatic capacitance force effect to the PFM amplitude image, where
panel (a) is what we suppose to see from a ferroelectric polarization induced
contrast and panel (b) is what we see when a strong capacitance force exists.
However, it can be reduced by placing the tip near the edge of sample surface
due to the small capacitive coupling.
In order to avoid the effect coming from attractive force and increase the
signal-to-noise ratio, a relatively large force constant (e.g. 40 N/m) tip should be
used in PFM imaging. With a “shorter” and “stronger” cantilever, the capacitive
force effect can be significantly reduced, and clear domain boundary feature can
be obtained in the PFM images.
comes from at least three integration force where only the first term is real piezo-
electric response. Electrostrictive and electrostatic responses should also be con-
sidered, especially when the piezoresponse is very small in those non-typical
ferroelectric materials (such as ultrathin films, 2D materials, and other emerg-
ing new materials in which only monolayer or a few layers of some 2D mate-
rials present weak ferroelectricity). To deeply understand the mechanism and
improve the PFM technology (such as resonance tracing; Xie et al. 2012) are very
important in PFM study of very thin films and new materials’ ferroelectricity and
domain structure.
The piezoresponse and PFM sensitivity can usually be enhanced by driving
the ac modulation voltage near the resonance of the cantilever-specimen system
as shown in Figure 4.14(a), and thus magnify the piezoresponse amplitude
A ∼ d33 V 0 Q by orders of magnitude, where Q is the quality factor of the system
Δω
V0eiox
Amplitude
d33V0Q ΔA
d33V0
d33V0ei(ox + φ)
ω0 ω′0 Frequency
7.3 150
6 Fit 100
amplitude
phase
4.7 50
Amplitude (a.u.)
Phase (°)
3.4 0
2.1 –50
0.8 –100
ω1 ω2
–0.5 –150
255 280 305 330 355
Frequency (kHz)
Figure 4.14 Schematics of dual frequency resonance tracking (DFRT) technique: (a)
tip-nanofiber harmonic oscillation system, (b) resonance enhancement using single
frequency, and (c) schematics of DFRT using actual experimental data of piezoresponse versus
ac driving frequency. Source: Xie et al. (2012). Reproduced with permission of The Royal
Society of Chemistry.
86 4 Ferroelectric Characterizations
typically in the range of 10–100 (Xie et al. 2012). For the single frequency
resonance enhancement technique, the frequency is locked at the resonance in
a particular location. As the resonance shifts during scanning due to the change
of contact stiffness, the lock-in will actually drive the oscillator away from the
resonance at the current location, resulting in loss of enhanced sensitivity,
as schematically shown in Figure 4.14b. The dual frequencies adopted, on
the other hand, allow us to trace resonance when it varies over the sample
surface and minimize the cross-talk with topography, as schematically shown
in Figure 4.14c.
15% LBFO 50 K
(002)
SRO(002)
10% LBFO 150 K
(002)
BFO
300 K
Figure 4.15 Typical X-ray 𝜃–2𝜃 scans of (a) BiFeO3 and La:BiFeO3 films with different La
concentrations on SrRuO3 /SrTiO3 /Si substrates, and (b) the (111) surface of [111]-poled
PMN-0.28PT single-domain crystal. Source: Adapted from Chu et al. (2008) and Li et al. (2016).
4.5 Domain Imaging and Polarization Mapping by Transmission Electron Microscopy 87
(0–13)c
4.05 ΔL (013)c 4.05 ΔL
4.00 (1–13)c ^
(113)c bc
4.00 (103)c 3.95
3.90 βL
3.95 30 °C 30 °C ^
ac
1.60 1.65 1.70 1.75
20
25
30
35
40
45
50
2.
2.
2.
2.
2.
2.
2.
(a) (b) (c)
^
cm ^ T = 200 °C (MA)
4.15 {103}c 4.20 {113}c
cc
{113}m {203}m
4.15
4.10 (–103)c (–1–13)c
(0–13)c 4.10
q⊥ (Å–1)
4.05 (–113)c
4.05 2ΔH
(1–13)c
4.00
(103)c ^ ^
4.00 bc am
(013)c 3.95 (113)c
βH
3.90
3.95 200 °C 200 °C βH
^
ac
1.60 1.65 1.70 1.75
20
25
30
35
40
45
50
^
2.
2.
2.
2.
2.
2.
2.
–bm
qII (Å–1) qII (Å–1)
(d) (e) (f)
Figure 4.16 Reciprocal space mapping of the MA and MC phases in PMN-xPT crystal. Source:
Ko et al. (2011). Reprinted with permission of Springer Nature.
higher temperature with the direction of the c-axis tilt toward [110]c. The high
temperature structural phase is similar to the normal R-BiFeO3 film in terms of
the direction of the distortion; however, it is apparently different in terms of the
large c-axis lattice parameter.
0.2 μm
(a) (b)
Figure 4.17 (a) Low magnification TEM images of the KNN single crystal along ⟨100⟩ axes
showing typical laminate domain structures and (b) the corresponding selected area electron
diffraction pattern. Source: Adapted from Chen et al. (2015).
–––
[001] [211] [401]
Figure 4.18 SAED patterns of different zone axes of Bi0.85 Nd0.15 FeO3 ceramics taken from the
of superstructure regions. The zone axes corresponding to (a–f ) are [210], [100], [010], [001],
[211] and [401], respectively. Source: Pu et al. (2015). Reprinted with permission of Elsevier.
4.5 Domain Imaging and Polarization Mapping by Transmission Electron Microscopy 89
20 pm
BFO 1
2
SRO
(b)
O
STO DFe
[001]
Fe
c
[100] a Bi
(a) (c)
Figure 4.19 (a) Atomically resolved HAADF-STEM images of 2 unit-cell-thick BiFeO3 . (b) B-site
atomic displacement vector map from the area marked with a white dashed rectangle in (a),
where the yellow arrows superimposed on the atomic image represent the relative
displacement DFe of the Fe atoms as shown in (c). Source: Adapted from Wang et al. (2018).
90 4 Ferroelectric Characterizations
Fe cation tetragonal cell that give rise to the spontaneous polarization. However,
due to difficulty in imaging the light element (O), the relative displacement
between the Fe cation and the mass center of an elongated rectangle formed by
its four Bi neighbors (denoted as DFe in Figure 4.19) can be used to determine
the polarizations of BFO unit cells.
Another important technique in TEM is the in situ observation with electric
field applied to the sample. Under electric field driven, domain evolution can be
dynamically observed (see Figure 4.20). No domain structure is observed without
electric field, but an applied positive electric field leads to the domain growth
Pup
(a) After contact (0 V) Nb-STO
(b) 3V
(c) 10 V
(d) –4 V
(e) –10 V
Figure 4.20 A series of TEM image of in situ observation of ferroelectric domain evolution
induced by external biases. Source: Han et al. (2014). Reprinted with permission of Springer
Nature.
4.5 Domain Imaging and Polarization Mapping by Transmission Electron Microscopy 91
toward the top electrode. When the negative field is applied, it is noted that the
domain configuration is reversible.
Exp. Sim.
c* Ba c*
001
Ti Ps
000 010 O
c
m b m
(a) (b) (c)
Figure 4.21 (a) CBED pattern of the tetragonal phase of PbTiO3 taken at room temperature
with the [100] incidence, where the mirror expected from the tetragonal structure is indicated
by a line. (b) Corresponding crystal structure of the tetragonal phase with space group P4mm.
(c) CBED pattern simulated with the structure in (b). Source: Tsuda and Tanaka (2013).
Reprinted with permission of IOP.
92 4 Ferroelectric Characterizations
References
Chen, Y., Wong, C.M., Dai, J.Y. et al. (2015). Ferroelectric domain structures in
⟨001⟩-oriented K0.15 Na0.85 NbO3 lead-free single crystal. AIP Advances 5 (3):
037117. Available at: https://aip.scitation.org/doi/pdf/10.1063/1.4914936?
class=pdf.
Cheng, Z.-Y., Katiyar, R.S., Yao, X., and Bhalla, A.S. (1998). Temperature dependence
of the dielectric constant of relaxor ferroelectrics. Physical Review B 57 (14):
8166–8177. Available at: https://link.aps.org/doi/10.1103/PhysRevB.57.8166.
Chu, Y.H., Zhan, Q., Yang, C.-H. et al. (2008). Low voltage performance of epitaxial
BiFeO3 films on Si substrates through lanthanum substitution. Applied Physics
Letters 92 (10): 102909. Available at: https://doi.org/10.1063/1.2897304.
Felten, F., Schneider, G.A., Saldaña, J.M., and Kalinin, S.V. (2004). Modeling and
measurement of surface displacements in BaTiO3 bulk material in piezoresponse
force microscopy. Journal of Applied Physics 96 (1): 563–568. Available at: https://
doi.org/10.1063/1.1758316.
Gruverman, A. and Kalinin, S.V. (2007). Piezoresponse force microscopy and recent
advances in nanoscale studies of ferroelectrics. In: Frontiers of Ferroelectricity
(eds. S.B. Lang and H.L.W. Chan), 107–116. Boston, MA: Springer US. Available
at: https://doi.org/10.1007/978-0-387-38039-1_10.
Han, M.-G., Marshall, M.S.J., and Zhu, Y. (2014). Interface-induced nonswitchable
domains in ferroelectric thin films. Nature Communications 5: 4693. Available at:
https://doi.org/10.1038/ncomms5693.
Hong, S., Shin, H., Woo, J., and No, K. (2002). Effect of cantilever-sample interaction
on piezoelectric force microscopy. Applied Physics Letters 80 (8): 1453–1455.
Ko, K.-T., Park, J.H., Yang, C.H. et al. (2011). Concurrent transition of ferroelectric
and magnetic ordering near room temperature. Nature Communications 2: 567.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/22127063.
Li, F., Zhang, S., Chen, L.Q. et al. (2016). The origin of ultrahigh piezoelectricity in
relaxor-ferroelectric solid solution crystals. Nature Communications 7: 13807.
Available at: http://dx.doi.org/10.1038/ncomms13807.
Liu, S., Kim, Y., Tan, L.Z., and Rappe, A.M. (2016). Strain-induced ferroelectric
topological insulator. Nano Letters 16 (3): 1663–1668. Available at: https://doi
.org/10.1021/acs.nanolett.5b04545.
Pu, S., Zheng, H., Zou, H. et al. (2015). The microstructure and ferroelectric
property of Nd-doped multiferroic ceramics Bi0.85 Nd0.15 FeO3 . Ceramics
International 41: 5498–5504.
Scott, J.F. (2008). Journal of Physics: Condensed Matter 20: 021001.
Tang, Y.L., Zhu, Y.L., Ma, X.L. et al. (2015). Observation of a periodic array of
flux-closure quadrants in strained ferroelectric PbTiO3 films. Science 348 (6234):
547–551. Available at: http://www.sciencemag.org/cgi/doi/10.1126/science
.1259869.
Tsuda, K. and Tanaka, M. (2013). Convergent-beam electron diffraction study of the
local structure of the tetragonal phase of PbTiO3 . Applied Physics Express 6 (10):
101501. Available at: http://stacks.iop.org/1882-0786/6/i=10/a=101501.
References 93
There are very broad applications for ferroelectric materials, especially thin
films, and the research interest in ferroelectrics is showing accelerated increase
in the twenty-first century. One of the main reasons is the successful demon-
stration of ferroelectric non-volatile memory in the field effect transistor (FET)
structure. Other new developments in ferroelectrics are stimulated by nanotech-
nology, where the ferroelectricity (FE) in nanometer scale and the corresponding
advanced characterization techniques for emerging new materials such as
two-dimensional (2D) and topological materials, have been extensively studied.
This chapter introduces the most recent progress in ferroelectric research
in terms of new materials, new phenomena, and advanced characterization
techniques and applications.
2 180
Height (nm)
Phase (°)
0 0
Figure 5.1 Demonstration of ferroelectricity for a 1-nm BaTiO3 film grown on top of a 30 nm
LSMO electrode. (a) Topography (636 mm2 ) and (b) PFM out-of-plane (OOP) phase image after
writing eight alternating positive (13.5 V) and negative (23.5 V) 430.5-mm2 voltage stripes. (c)
A 232-mm2 dark square is written on top of the pattern with 13.5 V, and a 131-mm2 white
square (23.5 V) is subsequently written on top of that square. Source: Garcia et al. (2009).
Reprinted with permission of Springer Nature.
materials are ferroelectric; while their bulk is not, since multilayers average out
each layer’s electric polarization.
Since the first prediction of 2D Van der Waals ferroelectric (Wu et al. 2013),
theoretical studies have proposed that 2D ferroelectrics may appear in distorted
monolayer MoS2 (Shirodkar and Waghmare 2014), buckled honeycomb AB
monolayer (DiSante et al. 2015), phosphorene analogues (GeS, GeSe, SnS,
and SnSe) (Wu and Zeng 2016), hydrogenated carbon nitride (Tu et al. 2017),
and chemically functionalized 2D materials (Wu et al. 2016). For instance,
first-principle calculations predict that the phosphorene analogue (GeS, GeSe,
SnS, and SnSe) monolayers are multiferroic with coupled ferroelectricity and fer-
roelasticity. The structure of SnS and SnSe monolayers are non-centrosymmetric,
and the parallel zigzag chains with dipole moments along the armchair direction
may lead to a spontaneous polarization. If the polarization is switchable, these
materials can be classified as intrinsic ferroelectrics.
Let’s have a look at SnSe or SnS as an example to understand its ferroelectric
nature. Figure 5.2 illustrates ferroelectric switching in SnS/SnSe monolayer. One
can see that when an electric field is applied to the right, the Sn atoms with pos-
itive charge will move toward the right from state I; while the S/Se atoms with
negative charge will move toward the left, then reach an intermediate state II with
a square lattice and finally get to the final state III with the polarization switched.
Therefore, this SnSe (or SnS) is intrinsically ferroelectric. It is interesting to note
that SnS/SnSe adopts a layered orthorhombic structure (space group Pnma) at
room temperature in state I originally. However, when the field is applied, the
structure converts to the nonpolar square structure (Cmcm phase) with the inver-
sion symmetry (intermediate state II) similar to the high temperature condition
(Li et al. 2015; Zhao et al. 2014), then back to the Pnma phase.
I ll Ill
P
Sn
S/Se
P
P
P
y
E-field direction
x
Figure 5.2 Illustrations of (top) top view and (bottom) side view along y direction of the
structure showing the pathway of ferroelectric switching for SnS/SnSe monolayer. Black
arrows denote external electric fields in different directions.
98 5 Recent Advances in Ferroelectric Research
0.5
0.0
100 200 300 400 500
(a) Temperature (K)
1.0 GeS
Polarization (C/m2)
0.8 GeSe
SnS
0.6 SnSe
0.4
0.2
0.0
1 3 5 7
(b) MX layers
5.3 Ferroelectric Vortex 99
Figure 5.4 (a) The stripe domain of a 1 unit cell SnTe film and (b) schematic of the lattice
distortion and atom displacement in the ferroelectric phase. Source: Adapted from Chang
et al. (2016).
Figure 5.5 Vortex structures in nature from flower (a) to snail (b), tornado (c), and galaxy (d).
The vortex domains are well studied in nanoferromagnets, but rarely reported
in ferroelectrics experimentally since it was extended to ferroelectrics around
1984 (Ginzburg et al. 1984) and predicted by Naumov and Fu (Naumov et al.
2004) from ab initio studies. Compared with magnetic vortex domain, the
ferroelectric vortex is more difficult to be formed due to its metastable state
and is also technologically difficult to be observed. Gruverman et al. (2008)
reported the direct observation of a doughnut-shaped domain pattern appearing
during polarization reversal in micrometer-sized ferroelectric capacitors (see
Figure 5.6). Domain imaging of the PZT disc was performed through the top
electrode of circular 1-μm diameter and square 2.5-μm edge by applying a read-
ing oscillating bias on a 100 ns scale. For the circular capacitors, an unswitched
domain remains in the center forming a vortex domain pattern. However, the
vortex structure remains only for >1 second, after that the unswitched central
domain collapses to leave a uniformly polarized ground state. Though it is
difficult to make unambiguous detection of the in-plane polarization, a 2D map
of the in-plane polarization distribution Pxy shows that the vortex structure is
consistent with the pattern as predicted by simulation, which is supported by
the model of Naumov and Fu.
Recently, Lane Martin and coworkers (Damodaran et al. 2017; Yadav et al.
2016) explored the coexistence of polarization vortices with other forms of polar
1 μs 3 μs Pxy
SrTiO3
PbTiO3
The curl of the polar displacement
[001]pc
5 nm
[100]pc
Figure 5.7 Exploring the phase boundary between (SrTiO3 )10 /(PbTiO3 )10 superlattice and
vortex–antivortex phases in each PbTiO3 layer. Source: Yadav et al. (2016). Reprinted with
permission of Springer Nature.
(a) (b)
Figure 5.8 (a) Schematic configurations of a single chiral skyrmion and (b) radial skyrmion
with the corresponding cross-section along the dashed line. The arrows indicate the direction
of the spins, and their colors represent the normal component to the plane, that is, from up
direction (red) to the down direction (blue). Source: Kezsmarki et al. (2015). Reprinted with
permission of Springer Nature.
m=1 ny
nx
m = –1
Figure 5.9 Skyrmion structures with varying vorticity m and helicity 𝛾. The arrows indicate the
direction of the in-plane order parameter component, and the brightness indicates the normal
component to the plane, with white denoting the up direction and black the down direction.
Source: Nagaosa and Tokura (2013). Reprinted with permission of Springer Nature.
The typical spin textures corresponding to a single skyrmion and radial skyrmion
are shown in Figure 5.8a and Figure 5.8b, respectively, each of them shows a vec-
tor field that all the vectors point along a given direction at the external surface
of the skyrmion.
Figure 5.9 shows a few patterns of the different possible tangent vector fields
in 2D xy-plane. Skyrmion structure can be indexed by vorticity m and helicity 𝛾,
where m stands for vorticity and measure how much the pattern rotates and 𝛾
indicates the initial angle that the vector field makes with the x-axis. With this
definition, the skyrmion in Figure 5.8a can be indexed as m = 1 and 𝛾 = 𝜋/2, and
the one shown in Figure 5.8b corresponds to m = 1 and 𝛾 = 𝜋.
By integrating ferroelectricity to magnetic skyrmion system, ferroelectrically
controlled magnetic skyrmions in ultrathin oxide heterostructures may be
formed. As reported by Wang et al. (2018), ferroelectric control of skyrmion
5.3 Ferroelectric Vortex 103
1.0
0.4 ρAHE + ρTHE
0.5
0.2
ρxy – R0H (μΩ cm)
M (μB/Ru)
0.0 0.0
–0.2
–0.5
ρAHE (M)
–0.4 B20S4 ρTHE
–1.0
–4 –2 0 2 4
μ0H (T)
Figure 5.10 Ordinary 𝜌xy , anomalous 𝜌AHE and topological Hall resistivities 𝜌THE versus H for
the BTO/SRO bilayer heterostructures. The 𝜌AHE –H curve is derived from the normalized M–H
curve. The unit of M is μB/Ru. The H-sweeping directions are marked by solid arrows. The inset
is the schematic skyrmion structure. Source: Adapted from Wang et al. (2018).
Phase (°)
Vtip = –8 V P +8 V +8 V +8 V
–8 V –8 V
0
ρxy
0.0
–0.2 Hsk
ρTHE + ρAHE
–0.4 ρTHE
Figure 5.11 (A) Schematic diagram of the experimental setup for FE domain switching and
Hall measurements. (B) PFM phase images (a), hall resistivity signals (b) of the BTO/SRO bilayer
in different FE poling states. The scale bar in (B) corresponds to 10 μm. The critical H as 𝜌THE
decreases to zero is marked as Hsk . Source: Adapted from Wang et al. (2018).
b
b α β
c
a c a
Figure 5.12 The structures of the two phases of DIPAB, α phase at 293 K and the β phase at
438 K, and the phase transition routes. Color scheme: C – black, N – cyan, H – white,
Br – golden. Green dashed bonds are N—H· · ·Br hydrogen bonds. Source: Adapted from Fu
et al. (2013).
large 𝜖 ′ up to 103 , and low tan 𝛿 of ∼0.4%, all close to or beyond the pure inorganic
BaTiO3 , representing a breakthrough in the research of molecular ferroelectrics.
Figure 5.12 shows the two phases 𝛼 and 𝛽 of DIPAB crystal. Below the Curie
temperature (T c = 426 K), the 𝛼 phase is in the monoclinic chiral space group P21 ,
which belongs to the polar point group C 2 . When heating the crystals to temper-
atures above T c , it results in the high temperature 𝛽 phase of a centrosymmetric
structure belonging to the monoclinic space group P21 /m in the nonpolar point
group C 2h . The 𝛼 to 𝛽 phase transition is reversible, and accompanied to the tran-
sition, the 𝛼 crystal displays a prominent dielectric anomaly at T c = 426 K, with
the peak 𝜖 ′ as high as 1.6 × 103 at the lowest frequency of 400 Hz (Figure 5.13a).
The polar 𝛼 crystal displayed typical ferroelectric hysteresis loops (Figure 5.13b,
inset). This result opens a new avenue to the applications of molecular ferro-
electrics that are much more environmental friendly, easier to fabricate, lighter
weight, and so on, compared with the pure inorganic oxide ferroelectrics.
1500 400 Hz 20
1 kHz 416 K 418 K
10 10
10 kHz
Ps (μC/cm2)
P (μC/cm2)
1 MHz
–10
500 –10
420 K 422 K
–10 0 10
E (kV/cm)
–20
0
380 400 420 440 380 350 400 450
(a) T (K) (b) T (K)
Figure 5.13 (a) The temperature dependence of the dielectric constant of the 𝛼-DIPAB crystal
under several frequencies. (b) The Ps versus T plots measured by a pyroelectric technique, and
inset, the ferroelectric hysteresis loops under several temperatures below T c . Source: Adapted
from Fu et al. (2013).
106 5 Recent Advances in Ferroelectric Research
Oxygen octahedron in
perovskite oxide
: Oxygen ion
: Metal ion
Oxygen hexahedron in
fluorite oxide c
a/b
Figure 5.15 A schematic for the structural origin of the ferroelectricity in perovskite and
fluorite ferroelectrics. Source: Adapted from Park et al. (2017).
tetragonal phase. Although t-phase and c-phase can be stabilized at room tem-
perature via doping or nanostructuring, all these structures are centrosymmetric
and reported as non-ferroelectric. However, the recent discovery of the possible
factor such as doping and crystallization with capping layer, HfO2 structure can
be stabilized into non-centrosymmetric orthorhombic phase (Pbc21 ) exhibiting
ferroelectricity (see Figure 5.16).
In 2011, Böscke et al. firstly observed a non-centrosymmetric orthorhombic
phase during the study of silicon doped HfO2 , which exhibited a remnant polar-
ization (Pr ) of ∼10 μC/cm2 at a coercive field of 1 MV/cm (see Figure 5.17). In fact,
ferroelectricity (FE) or antiferroelectric (AFE) can be induced in HfO2 by several
T phase, P42/nmc
20
2.6 mol%
15
3.1 mol%
4.3 mol%
10
5.6 mol%
Polarization (μC/cm2)
–5
–10
–15
–20
–2 –1 0 1 2
Field (MV/cm)
Table 5.1 The ferroelectric properties of HfO2 films with various dopants (O: observable, X: not
observable).
Si (4+) 41 24 0.8–1.0 4 O
Zr (4+) 50 34 5 50 X
Y (3+) 40 24 1.2–1.5 5.2 X
Al (3+) 48 16 1.3 4.8 O
Gd (3+) 53 12 1.75 2 X
Sr (2+) 48 23 2.0 9.9 X
La (3+) N/A 45 1.2 N/A N/A
Source: Adapted from Park et al. (2015) and Wei et al. (2018).
dopants (Y, Gd, Al, Sr, La, Mg, Ba, Nd, Sm, Er, Ga, In, Co, N, and Ni), especially
for ZrO2 -doped HfO2 . As shown in Table 5.1, the largest reported value of Pr
is 45 μC/cm2 , observed in La-doped HfO2 film. On the other hand, researchers
have succeeded to suppress the monoclinic phase in sub-10 nm HfO2 by con-
trolling the oxygen fraction in deposition ambient, and a Pr of 13.5 μC/cm2 in a
6.9-nm-thick dopant-free HfO2 film has been observed by Ashish Pal et al. (2017).
5.5 Ferroelectricity in HfO2 and ZrO2 Fluorite Oxide Thin Films 109
HfO2 has a large bandgap (Eg > 5 eV), a larger remnant polarization
(Pr ∼ 45 μC/cm2 ) and coercive field (∼1–2 MV/cm) than conventional ferro-
electric films. The recent discovery of ferroelectricity in HfO2 , which only
appears at the nanoscale and becomes stronger at smaller dimensions, makes
it able to retain ferroelectricity for extremely thin films (<10 nm). Also, the
simple binary oxide structure with high dielectric constant (high-k), silicon
compatibility, and low toxicity makes it very suitable for applications as a gate
dielectric in semiconductor MOS-FET for ferroelectric random-access memory
(FeRAM) application, which has been introduced in Chapter 3. FE/AFE-HfO2
film shows promising applications in ferroelectric random access memory,
pyroelectric energy harvesting and storage.
It is interesting to note that the ferroelectricity of HfO2 only appears at
nanoscale. For the conventional ferroelectrics, the ferroelectric properties in
thin film form can only be retained under the compressive clamping effect with
the substrate. By contrast, remnant polarizations of pure or doped HfO2 is
larger than 15 μC/cm2 when thickness is 10 nm or even thinner (Park et al. 2013,
2014; Wei et al. 2018; Polakowski and Müller 2015). For instance, Polakowski
and Müller (2015) show the P–V and C–V hysteresis loops for a thickness
series of pure HfO2 thin films crystallized in the presence of a PVD-TiN top
electrode (Figure 5.18). It shows the degrading of ferroelectricity when the film
becomes thicker. To explain it, they claimed that the gradual destabilization of
the presumably orthorhombic phase in favor of the monoclinic phase is observed
with increasing thickness.
20
10
Polarization (μC/cm2)
20 nm
12 nm
–10 10 nm
8 nm
6 nm
–20
–2.5 –2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 2.5
Field (MV/cm)
Figure 5.18 P–E hysteresis loops for a thickness series of pure HfO2 thin films crystallized in
the presence of a PVD-TiN top electrode. Source: Adapted from Polakowski and Müller (2015).
110 5 Recent Advances in Ferroelectric Research
Several researchers have also studied the phase evolution in HfO2 film with
different thicknesses (Park et al. 2013, 2014, 2017; Wei et al. 2018). The following
evolution trends as a function of thickness are summarized by the experimental
result of grazing incidence X-ray diffraction (GIXRD) and transmission electron
microscopy (TEM) observations. In ultrathin film (thickness < 10 nm), the
amount of o-phase is the highest, and m-phase is absent or very rarely found.
When the thickness increases, the crystal size (grain size) grows further, the
m-phase starts forming with relieving the internal stress. Once the film is
significantly thick (thickness > 20 nm), the m-phase crystal structure appears as
the predominant phase. Thus, the effective electric field applied to ferroelectric
phase is decreased as a consequence of the increase amount of non-ferroelectric
phase. It can therefore explain the generally decreasing polarization with the
increasing thickness.
Since the scaling of the physical thickness of ferroelectric thin film is the
fundamental issue for FeRAM fabrication, the thickness of the conventional
perovskite-structure ferroelectrics is limited to achieve robust ferroelectricity.
The remnant polarization of perovskite oxide is not greater than that of fluorite
oxide, and the difficulty of depositing perovskite-structure ferroelectrics with
two or more cation components in high-aspect-ratio structure via atomic layer
deposition (ALD) is a critical issue. This makes fluorite-structure ferroelectric
able to replace conventional ferroelectrics for memory applications.
However, in 2013, Zhou et al. discovered a critical problem, the so-called
“wake-up effect,” in Si-doped HfO2 ferroelectric film. This effect refers to the
increase in remnant polarization with increasing cycling number of switching
field. Unlike conventional ferroelectrics such as BaTiO3 and PZT, the ferroelec-
tricity of HfO2 relies on large number of switching cycles, making it difficult for
application. Figure 5.19 shows the general switching I–E curves and remnant
polarization change of ferroelectric HfO2 -based thin films with three different
states of field cycling. From the graph, one can see that the three states, starting
Wake-up Fatigue
Pr Wake-up
Pr,rel Fatigue
Polarization (a.u.)
Pr /Pr,relax (a.u.)
Current (a.u.)
Pristine
Figure 5.19 (a) Evolution of the ferroelectric remnant polarization during bipolar cycling (4 V
at 10 kHz) measured on Sr:HfO2 -based ferroelectric capacitors. (b) Current–voltage
characteristics for three states, pristine, wake-up, and fatigue. Source: Park et al. (2018).
Reprinted with permission of Cambridge University Press.
5.5 Ferroelectricity in HfO2 and ZrO2 Fluorite Oxide Thin Films 111
from pristine, wake-up to fatigue, can be observed during the cycling. The
remnant polarization as a function of cycles shows two distinct trends as
described by Pešić et al. (2016):
i. “Wake-up” – increases with cycling. This increase of the Pr value corresponds
to an opening or de-pinching of the pristine pinches hysteresis loop with field
cycling.
ii. “Fatigue” – the Pr value decreases resulting in a closure of the difference
between positive and negative remnant polarizations after a certain number
of cycles due to aging mechanisms.
As shown in the current–voltage characteristics curve, the switching current
curve has double peaks in both positive and negative bias regions in pristine
state and, consequently, forms a pinched hysteresis loop. It is different from
the ideal switching current curve of a ferroelectric, where only a single peak is
seen at each region. However, the switching current peaks move toward each
other with continuous bipolar cycling and finally merge together in the wake-up
regime. When cycling further increases, the switching polarization decreases,
which corresponds to the typically observed fatigue in many ferroelectrics.
According to the discussion of Pešić et al. (2016) and Schenk et al. (2014),
the behavior of splitting of switching current peaks in the pristine state may
come from different built-in field within the film. The inhomogeneous internal
electric field distribution can originate from polycrystallinity of the film and the
unevenly distributed charged defects, such as oxygen vacancies. A high oxygen
vacancy concentration is expected to be present near the electrodes due to
the reduction of the doped HfO2 layer by the metal nitride electrodes, whose
distribution may be mostly asymmetric. Therefore, the resulting asymmetric
distribution of the oxygen vacancies can be an origin of the internal field in
the pristine material. When it is in cycling process, the oxygen vacancies at
the interface may diffuse into the bulk regions of the ferroelectric-HfO2 -based
films, and their distribution becomes more homogeneous, in which the wake-up
process will occur, accompanied by the merging of the switching current
peaks.
Another mechanism of the wake-up effect is discussed as the field cycling-
induced phase transition. From the observation of a decrease in the dielectric
constant and an increase in remnant polarization with an increasing number
of cycling, a phase transition from a non-ferroelectric phase to a ferroelectric
phase was suggested in 2015 (Lomenzo et al. 2015). Later on, many experimen-
tal results show both direct and indirect evidences suggesting a phase transi-
tion to the ferroelectric o-phase from non-ferroelectric phase, either t-phase or
m-phase. In Figure 5.20, the phase evolution trend shows the schematic tran-
sition to o-phase in pristine, wake-up, and fatigue states, which indicates that
oxygen vacancy distribution as well as the phase transition plays a significant role
in fluorite-structure ferroelectric wake-up process.
Owing to the industrial-compatible fabrication process, most of the
ALD-grown HfO2 films are polycrystalline and contain multiple phases.
That is the reason why a wake-up effect and further treatment have to be done
before obtaining the better ferroelectricity. Recently, highly oriented ferroelectric
112 5 Recent Advances in Ferroelectric Research
TiN
Mix phase
TiOxNy
O-p O- O-
has ph ph
e as as
Vo e e
M- M-
M-p ph ph
has as as
e Sr;HfOx e e
Mixed phase
TiO2
TiN
(a) (b) (c)
Figure 5.20 Phase evolution model of the three different stages of the ferroelectric
capacitance lifetime. (a) Prestine, (b) wake-up, and (c) fatigue. Source: Adapted from Pešić et al.
(2016).
[111]
[111]
[111] [11-1]
[11-1] [11-1]
LSMO (φ = 15°)
2 nm
(a)
b
[111] a
[-111]
[11-1] ~71° [1-11]
120°
a b
c
(b) (c)
Figure 5.21 (a) Cross-sectional high-angle annular dark field-scanning transmission electron
microscopy (HAADF-STEM) image of a 4-nm-thick film, observed along the zone [111], (b)
sketch of the proposed rhombohedral structure of the HZO film with polarization along the
[111] direction, and (c) two views of the R3m phase obtained for epitaxially compressed HZO
and HfO2 . Green (cyan) spheres represent Hf/Zr (O) atoms. The arrows show the polar
distortion with respect to a reference paraelectric structure with appropriately R3m state.
Source: Wei et al. (2018). Reprinted with permission of Springer Nature.
300 OFF-field
90
250
PFM amplitude (a.u.)
Column-like 200
0
AI:HfO2 150
100
50 –90
LSMO
10 nm [111] 0
–20 –10 0 10 20
Voltage (V)
(a) (b)
Figure 5.22 (a) TEM image of Al:HfO2 /LSMO and (b) PFM amplitude and phase hysteresis
loops from doped HfO2 on STO (111) substrate.
114 5 Recent Advances in Ferroelectric Research
Table 5.2 Calculated properties of four hybrid lead halide perovskites from
density functional theory.
NH4 6.21 8
CH3 NH3 6.29 38
CF3 NH3 6.35 48
NH2 CHNH2 6.34 63
500 200
400
150
Amplitude (a.u.)
50
Phase (°)
300
0 100
200
–50 50
100
200 nm –100
0
0
–4 –3 –2 –1 0 1 2 3 4
(a) (b)
Voltage (V)
Figure 5.23 (a) Surface morphology of CH3 NH3 PbI3 perovskite crystal showing strip-like
domain structure and (b) piezoelectric butterfly loop of CH3 NH3 PbI3 perovskite crystal
(J.Y. Dai 2013 unpublished work).
by Grinberg et al. (2013). Our group has been studying the existence of ferro-
electricity in CH3 NH3 PbI3 single crystal for many years. Figure 5.23 shows the
AFM/PFM results of CH3 NH3 PbI3 single crystal where the surface presents
domain structure and PFM strain-bias butterfly loop. These results suggest that
CH3 NH3 PbI3 single crystal may have ferroelectric property.
Our group has tried to combine ferroelectric material into this perovskite
solar cell structure as shown in Figure 5.24. We proposed structures utilizing fer-
roelectric nanodots embedded in the CH3 NH3 PbI3 perovskite or a ferroelectric
network structure at the bottom and top electrodes. It is expected that polarized
Polarized discrete
ferroelectric layer
Polarized discrete D Polarized discrete
TA
AD
ferroelectric layer
3
ferroelectric layer
bl
3
bl
P Me
eT
3P
3P
Au
Au
M
-O
H
H
-O
3N
iro
3N
iro
Sp
H
P
H
P
Sp
C
C
O
O
Zn
O
Zn
ss
IT
IT
ss
la
G
la
G
(b)
(a)
AD
3
eT
bl
3P
M
Au
-O
H
iro
3N
Sp
H
C
P
Zn
Ferroelectric nanodots
IT
ss
la
G
Figure 5.24 (a) A discrete ferroelectric layer in between electron-collection layer and ITO
electrode, (b) double ferroelectric layers at cathode and anode, (c) ferroelectric nanodots
embedded 0–3 nanocomposite perovskite solar cell structure, and (d) hexagonal nanocells of
PZT discrete film on ITO/glass made by sol–gel method followed by nano-patterning
technique. The conductivity of the ITO layer underneath has negligible degradation during the
PZT synthesis process (J.Y. Dai and H.J. Fang 2014 unpublished).
116 5 Recent Advances in Ferroelectric Research
pm
250
0.0
2 μm 2 μm
–250
(a) (b)
a.u.
545
310
4 μm 2 μm
75.0
(c) (d)
Figure 5.25 Ferroic domain patterns of CH3 NH3 PbI3 crystals revealed by (a) SEM, (b) AFM
topography, (c) polarized optic microscope, and (d) vertical PFM amplitude. Source: Adapted
from Huang et al. (2018).
References 117
suggests that those alternating domains have their different origins. It is found
that polar and nonpolar orders of CH3NH3PbI3 can be distinguished from their
distinct lateral piezoresponse, energy dissipation, first and second harmonic
electromechanical couplings, and temperature variation, even though their dif-
ference in crystalline lattice is very subtle. These findings resolve key questions
regarding polar nature and ferroelectricity of CH3NH3PbI3 and its implication
on photovoltaics.
References
Böscke, T.S., Müller, J., Bräuhaus, D. et al. (2011). Ferroelectricity in hafnium oxide
thin films. Applied Physics Letters 99 (10): 102903. Available at: https://doi.org/10
.1063/1.3634052.
Chang, K., Liu, J., Ji, S.H. et al. (2016). Discovery of robust in-plane ferroelectricity
in atomic-thick SnTe. Science 353 (6296): 274–278. Available at: http://science
.sciencemag.org/content/353/6296/274.
Cheetham, A.K. and Rao, C.N.R. (2007). There’s room in the middle. Science 318
(October): 58–60.
Dai, J.Y., Lee, P.F., Wong, K.H. et al. (2003). Epitaxial growth of yttrium-stabilized
HfO2 high-k gate dielectric thin films on Si. Journal of Applied Physics 94 (2):
912–915. Available at: https://doi.org/10.1063/1.1585116.
Dall’Olio, S. and Dovesi, R. (1997). Spontaneous polarization as a Berry phase of the
Hartree–Fock wave function: the case of KNbO3 . Physical Review B: Condensed
Matter 56 (16): 10105–10114.
Damodaran, A.R., Clarkson, J.D., Martin, L.W. et al. (2017). Phase coexistence and
electric-field control of toroidal order in oxide superlattices. Nature Materials 16:
1003. Available at: http://dx.doi.org/10.1038/nmat4951.
DiSante, D., Stroppa, A., Barone, P. et al. (2015). Emergence of ferroelectricity and
spin-valley properties in two-dimensional honeycomb binary compounds.
Physical Review B: Condensed Matter 91 (161401(R)): 1–6.
Fei, R., Kang, W., and Yang, L. (2016). Ferroelectricity and phase transitions in
monolayer group-IV monochalcogenides. Physical Review Letters 117 (097601):
1–6.
Fong, D.D., Stephenson, G.B., Streiffer, S.K. et al. (2004). Ferroelectricity in ultrathin
perovskite films. Science 304 (5677): 1650–1653. Available at: http://science
.sciencemag.org/content/304/5677/1650.
Frost, J.M., Butler, K.T., Brivio, F. et al. (2014a). Atomistic origins of
high-performance in hybrid halide perovskite solar cells. Nano Letters 14 (5):
2584–2590. Available at: https://doi.org/10.1021/nl500390f.
Frost, J.M., Butler, K.T., and Walsh, A. (2014b). Molecular ferroelectric
contributions to anomalous hysteresis in hybrid perovskite solar cells. APL
Materials 2 (8): 81506. Available at: https://doi.org/10.1063/1.4890246.
Fu, D.-W., Zhang, W., Xiong, R.G. et al. (2011). Diisopropylammonium chloride: a
ferroelectric organic salt with a high phase transition temperature and practical
utilization level of spontaneous polarization. Advanced Materials 23 (47):
5658–5662. Available at: http://dx.doi.org/10.1002/adma.201102938.
118 5 Recent Advances in Ferroelectric Research
Landau, L.D. and Lifshitz, E.M. (1935). On the theory of the dispersion of magnetic
permeability in ferromagnetic bodies. Physikalische Zeitschrift der Sowjetunion 8:
153–164.
Lee, P.F., Dai, J.Y., Wong, K.H. et al. (2003). Growth and characterization of
Hf–aluminate high-k gate dielectric ultrathin films with equivalent oxide
thickness less than 10 Å. Journal of Applied Physics 93 (6): 3665–3667. Available
at: https://doi.org/10.1063/1.1554764.
Li, C.W., Hong, J., Delaire, O. et al. (2015). Orbitally driven giant phonon
anharmonicity in SnSe. Nature Physics 11: 1063. Available at: http://dx.doi.org/10
.1038/nphys3492.
Lines, M.E. and Glass, A.M. (2001). Principles and Applications of Ferroelectrics and
Related Materials. Oxford: Oxford University Press. Available at: http://www
.oxfordscholarship.com/10.1093/acprof:oso/9780198507789.001.0001/acprof-
9780198507789.
Lomenzo, P.D., Takmeel, Q., Nishida, T. et al. (2015). TaN interface properties and
electric field cycling effects on ferroelectric Si-doped HfO2 thin films. Journal of
Applied Physics 117 (13): 134105.
Long, J.R. and Yaghi, O.M. (2009). The pervasive chemistry of metal-organic
frameworks. Chemical Society Reviews 38 (5): 1213–1214. Available at: http://dx
.doi.org/10.1039/B903811F.
Materlik, R., Kunneth, C., and Kersch, A. (2015). The origin of ferroelectricity in
Hf1−x Zrx O2 : a computational investigation and a surface energy model. Journal of
Applied Physics 117 (13): 134109.
Mermin, N.D. (1979). The topological theory of defects in ordered media. Reviews of
Modern Physics 51 (3): 591–648.
Nagaosa, N. and Tokura, Y. (2013). Topological properties and dynamics of
magnetic skyrmions. Nature Nanotechnology 8 (12): 899–911. Available at: http://
www.ncbi.nlm.nih.gov/pubmed/24302027.
Naumov, I.I., Bellaiche, L., and Fu, H. (2004). Unusual phase transitions in
ferroelectric nanodisks and nanorods. Nature 432: 737. Available at: http://dx.doi
.org/10.1038/nature03107.
Pal, A., Narasimhan, V.K., Weeks, S. et al. (2017). Enhancing ferroelectricity in
dopant-free hafnium oxide. Applied Physics Letters 110 (2): 22903. Available at:
https://doi.org/10.1063/1.4973928.
Park, M.H., Kim, H.J., Hwang, C.S. et al. (2013). Evolution of phases and
ferroelectric properties of thin Hf0.5 Zr0.5 O2 films according to the thickness and
annealing temperature. Applied Physics Letters 102 (24): 242905.
Park, M.H., Kim, H.J., Hwang, C.S. et al. (2014). The effects of crystallographic
orientation and strain of thin Hf0.5 Zr0.5 O2 film on its ferroelectricity. Applied
Physics Letters 104 (7): 072901.
Park, M.H., Lee, Y.H., Hwang, C.S. et al. (2015). Ferroelectricity and
antiferroelectricity of doped thin HfO2 -based films. Advanced Materials 27 (11):
1811–1831. Available at: http://dx.doi.org/10.1002/adma.201404531.
Park, M.H., Schenk, T., Schroeder, U. et al. (2017). A comprehensive study on the
structural evolution of HfO2 thin films doped with various dopants. Journal of
Materials Chemistry C 5: 4677–4690.
120 5 Recent Advances in Ferroelectric Research
Park, M.H., Schroeder, U., Hwang, C.S. et al. (2018). Review and perspective on
ferroelectric HfO2 -based thin films for memory applications. MRS
Communications 8 (3): 795–808.
Pešić, M., Fengler, F.P.G., Schroeder, U. et al. (2016). Physical mechanisms behind
the field-cycling behavior of HfO2 -based ferroelectric capacitors. Advanced
Functional Materials 26 (25): 4601–4612.
Polakowski, P. and Müller, J. (2015). Ferroelectricity in undoped hafnium oxide.
Applied Physics Letters 106 (23): 232905.
Rao, C.N.R., Cheetham, A.K., and Thirumurugan, A. (2008). Hybrid
inorganic–organic materials: a new family in condensed matter physics. Journal
of Physics: Condensed Matter 20 (8): 083202. Available at: http://stacks.iop.org/
0953-8984/20/i=8/a=083202.
Resta, R. (1994). Macroscopic polarization in crystalline dielectrics: the geometric
phase approach. Reviews of Modern Physics 66 (3): 899–915.
Schenk, T., Yurchuk, E., Mueller, S. et al. (2014). About the deformation of
ferroelectric hystereses. Applied Physics Reviews 1 (4): 41103. Available at: https://
doi.org/10.1063/1.4902396.
Shaw, T.M., Trolier-McKinstry, S., and McIntyre, P.C. (2000). The properties of
ferroelectric films at small dimensions. Annual Review of Materials Science 30:
263.
Shirodkar, S.N. and Waghmare, U.V. (2014). Emergence of ferroelectricity at a
metal-semiconductor transition in a 1T monolayer of MoS2 . Physical Review
Letters 112 (157601): 1–5.
Skyrme, T.H.R. (1962). A unified field theory of mesons and baryons. Nuclear
Physics 31: 556–569. Available at: http://www.sciencedirect.com/science/article/
pii/0029558262907757.
Stranks, S.D., Eperon, G.E., Snaith, H.J. et al. (2013). Electron-hole diffusion lengths
exceeding 1 micrometer in an organometal trihalide perovskite absorber. Science
342 (6156): 341–344. Available at: http://science.sciencemag.org/content/342/
6156/341.
Strelcov, E., Dong, Q., Li, T. et al. (2017). CH3 NH3 PbI3 perovskites: ferroelasticity
revealed. Science Advances 3 (4): e1602165. Available at: http://advances
.sciencemag.org/content/3/4/e1602165.
Tu, Z., Wu, M., and Zeng, X.C. (2017). Two-dimensional metal-free organic
multiferroic material for design of multifunctional integrated circuits. Journal of
Physical Chemistry Letters 8 (9): 1973–1978. Available at: https://doi.org/10.1021/
acs.jpclett.7b00636.
Wang, L., Lu, Q., Noh, T.W. et al. (2018). Ferroelectrically tunable magnetic
skyrmions in ultrathin oxide heterostructures. Nature Materials 17 (December):
1087–1094. Available at: http://dx.doi.org/10.1038/s41563-018-0204-4.
Wei, Y., Nukala, P., Noheda, B. et al. (2018). A rhombohedral ferroelectric phase in
epitaxially strained Hf0.5 Zr0.5 O2 thin films. Nature Materials 17 (December):
1095–1101. Available at: http://dx.doi.org/10.1038/s41563-018-0196-0.
Williams, W. and Wright, T.M. (1998). High-resolution micromagnetic models of
fine grains of magnetite. Journal of Geophysical Research, B: Solid Earth 103
(B12): 30537–30550. Available at: http://dx.doi.org/10.1029/98JB01120.
References 121
+ + + +
Piezoelectric
plate P
– – – –
Electrodes
stress. The following sections will give a detailed description of all parameters in
piezoelectric materials.
each constant should have two subscripts indicating the directions of the two
related quantities, electric field and strain, for piezoelectricity. The direction of
positive polarization is usually made to coincide with the z-axis of a rectangular
system of x-, y-, and z-axes (Figure 6.2). Directions of x, y, and z are represented
by the subscripts 1, 2, and 3, respectively, and shear about one of these axes is
represented by the subscripts 4, 5, and 6, respectively.
Piezoelectric constant is an important indicator of a material’s suitability
for strain-dependent applications. The following give definition of the most
frequently used physical constants and their interrelating for piezoelectric
materials, such as piezoelectric charge constant d, the piezoelectric voltage
constant g, and the dielectric permittivity 𝜖.
Piezoelectric plate
Electrodes
Polarization direction
(a) (b) (c) Stress direction
Figure 6.3 Diagram showing the definitions of (a) d33 , (b) d31 , and (c) d15 .
126 6 Piezoelectric Effect: Basic Theory
field applied in direction 3. In this case, the applied electric field or stress should
be parallel to direction in which piezoelectric plate is polarized.
d31 is similar to the definition of d33 , but for d31 , the applied stress or the induce
strain should be along the direction 1, and the measured polarization or the
applied E-field should be in direction 3. In this case, the applied electric field
or stress should also be parallel to the direction in which piezoelectric plate is
polarized.
d15 is the induced shear strain in direction 5 per unit electric field applied in direc-
tion 1 or the sample’s polarization generated along direction 1 by a shear stress
applied in direction 5. In this case, the applied electric field or stress should be
perpendicular to direction in which piezoelectric sample is polarized.
There are two contributions to the generated polarization Pi , i.e. the applied
stress and electric field. With the absence of an applied field, the polarization
produced by a stress 𝜎 jk , which is a second-rank tensor, is
Pi = 𝜖0 𝜒ij Ej (6.3)
where 𝜒 ij denotes the dielectric susceptibility measured at constant stress.
By combining these two contributions, the generated polarization to a linear
piezoelectric material can be expressed as
It should be noted that this expression is only right at low field regime where
the linear relationship is hold. One should also note that dijk is a third-rank ten-
sor, but by invoking elastic symmetries along with electric symmetries due to
poling, the tensor can be reduced to second rank, i.e. what has been introduced
as dij , typically d33 , d31 , and d15 . In fact, d31 = d32 ; and d15 = d24 , and all others
are zero.
For the converse piezoelectric effect, the strain is expressed as
𝜀ij = Sijkl
E
𝜎kl + dkij Ek (6.5)
applications, and the very often used g 33 , g 31 , and g 15 are defined in the same way
as Figures 6.2 and 6.3, i.e.
( ) ( )
𝜕Ei E 𝜕𝜀j 𝜎
gij = − = (6.6)
𝜕𝜎j 𝜕Di
where the “−” sign is due to the fact that the field built from electric polarization
reduces under positive compressive stress.
Dielectric permittivity is the linkage between the piezoelectric charge and volt-
age constants. This can be understood by the relation between voltage and charge,
whereas their ratio is capacitance, which is proportional to permittivity. There-
fore, the relationship between d and g is
d
g= (6.7)
𝜖0 𝜖r
From the relationship of d and g, we can understand why PVDF is a good choice
for sensor applications such as hydrophone and pressure sensor. It is the very low
dielectric constant that gives rise to a larger g factor, even its d factor is much
smaller than PZT; while PZT is excellent for transducer application due to its
large d value, making it able to generate large strain under electric field.
ε1 E
S11 E
S12 E
S13 0 0 0 0 0 d31 σ1
ε2 E E E
S12 S22 S13 0 0 0 0 0 d31 σ2
ε3 E
S13 E
S13 E
S33 0 0 0 0 0 d33 σ3
ε4 0 0 0 E
S44 0 0 0 d15 0 σ4
ε5 = 0 0 0 0 E
S44 0 d15 0 0 σ5
ε6 0 0 0 0 0 E
S66 0 0 0 σ6
σ
D1 0 0 0 0 d15 0 ϵ11 0 0 E1
σ
D2 0 0 0 d15 0 0 0 ϵ 11 0 E2
D3 d31 d31 d33 0 0 0 0 0 ϵ σ33 E3
It summarizes the converse and direct piezoelectric effects. In matrix form, the
components can be written as
𝜀 = SE 𝜎 + d∗ E
D = d𝜎 + 𝜖 𝜎 E
where d* denotes the transpose of d.
Note: With all the previously mentioned definitions, strain and dielectric con-
stant share the same letter 𝜀 as in most literatures, i.e. 𝜀 is used to represent strain
and dielectric constant. For friendly reading, in this book, 𝜀 and 𝜖 denote strain
and dielectric constant, respectively.
Depending on the vibration modes of shapes and the directions of electric field
and polarization, the electromechanical coupling factors can be defined as k 33 ,
k 31 , k 15 , k t , and k p . Detailed information about the requirement of testing samples
to satisfy these vibration modes will be introduced in later sections.
Poling direction
k33
k15
kp, kt
1 k31 Electrodes
Figure 6.4 Diagram showing the definition (the relationship between the poling and electric
field direction, where the electric field direction is parallel to the two electrodes) of k33 , k31 , k15 ,
kt , and kp.
130 6 Piezoelectric Effect: Basic Theory
Under Static or Low Frequency For a rod/bar sample with 33 vibration mode,
2
2
d33
k33 = E 𝜎
S33 𝜖33
For a plate sample with 31 vibration mode,
2 D
d31 S11
2
k31 = E 𝜎
=1− .
S11 𝜖33 E
S11
For a plate sample with 15 vibration mode,
2
2
d15
k15 = E 𝜎
S15 𝜖11
For a disc sample with thickness vibration mode,
2 𝜀
d33 𝜖33 2
d33
kt2 = = E 𝜎
D
C33 S33 𝜖33
For a disc sample with radial vibration mode (Meitzler et al. 1973),
2 2
2(k P )2 2k31 2d31
kp2 = = = 𝜎
,
1 + υp + 2(k P )2 1 − υp 𝜖33 E
(S11 E
+ S12 )
SE
where υp = − S12E as a planar Poisson’s ratio, and a planar radial piezoelectric cou-
11
p 2
(e31 )2 p 𝜎 2d31
pling coefficient k P is given by (k P )2 = p p
𝜖33 C11
,where 𝜖33 = 𝜖33 − E
S11 E
+S12
,
E
p S11
C11 = E 2 E 2
(S11 ) − (S12 )
p d31
e31 = E E
S11 + S12
k31 = √
√ [ ]
1 + 𝜋2 fa tan 𝜋2 a f r
f (f −f )
r r
2(k P )2
kp2 =
1 + υp + 2(k P )2
where
E
S12
𝜐p = − E
S11
2
fa2 − fr2
keff =
fa2
fr (fm)
Frequency f
132 6 Piezoelectric Effect: Basic Theory
Cp
90
fa
B
Impedance Z (Ω)
100
Phase ϕ (º)
E
A C
0
10
fr
D
–90
1
3.0 3.5 4.0 4.5 5.0 5.5 6.0
Frequency f(MHz)
Figure 6.7 Impedance-frequency spectrum of a poled PMN–PT single crystal with f r and f a .
A and E C
XL XL
R R
Cs Rs Rs Ls
Xc Xc
B XL D Rp YL
Ls Rs Cs 1
R
Rp
Lp
Xc Yc
Cp
where YL = 1 = 1
XL jωL
and YC = 1 = jωC
XC
Figure 6.8 Simplified equivalent circuit corresponding to the regions A–E in Figure 6.7.
In the equivalent circuit, impedance Z and its phase angle 𝜑 depend upon the
reactance values of the components; and we know that reactance X is zero when
the circuit element is resistive, positive when the circuit element is inductive, and
negative when it is capacitive. Their impedances are quantified:
ZR = R
1
ZC =
j𝜔C
ZL = j𝜔L
where j is the imaginary notation. In the previously mentioned equations, it is
apparent that the impedance of an inductor is directly proportional to the fre-
quency, while that of capacitor is inversely proportional to frequency. As a result,
134 6 Piezoelectric Effect: Basic Theory
1
2𝜋f Ls − =0
2𝜋f Cs
1
f = fr = √
2𝜋 Ls Cs
When the frequency further increases, the circuit element becomes inductive
that X L is larger than X C . Thus, it becomes inductive and phase angle is positive
(i.e. region C).
In region D, the piezoelectric material behaves like a parallel LC circuit. As the
frequency increases, the parallel circuit will produce a parallel resonance (also
called anti-resonance) circuit when the resultant current through the parallel
combination is in phase with the applied voltage. At anti-resonance, there is a
large circulating current between the inductor and the capacitor due to the energy
of the oscillations, and then parallel circuit produces current resonance. Dissim-
ilar to region B, the parallel LC circuit acts like an open circuit rather than short
circuit, so the current flow into the circuit is zero at anti-resonance.
For a parallel equivalent circuit, admittance Y can better illustrate the situation,
which is given as
1 1
Y = − j𝜔Cp +
RP j𝜔LP
where at anti-resonance state, the admittances of L and C are the same values
but antiphase resulting in a minimum value of total admittance. Thus, the total
impedance of this resonance circuit at anti-resonance is at its maximum value
creating a circuit condition of high resistance and low current. At f = f a , the
impedance of the circuit represents the so-called maximum dynamic impedance,
i.e. Zd = Z max .
When parallel resonance occurs, i.e. Z is maximum, and the imaginary parts of
Y become zero,
Im(Y ) = 0
so
1
− 2𝜋f LP = 0
2𝜋f Cp
6.4 Characterization of Piezoelectric Resonator Based on a Resonance Technique 135
Shape Dimensions Poling direction Vibration mode Resonant mode Related parameters
l – length, Piezoelectric Mechanical
w – width,
t – thickness,
d – diameter
𝜎
t > 10d Along t Thickness extension Thickness kt , 𝜖33 D
C33 , C33
E
, S33
E
, S33
D
, Qt
mode resonance
𝜀
l > 5(t, w) Along t Transverse length Length extensional k31 , d31 , g31 , 𝜖33 D
S11 , S11
E
, Q31
mode resonance
𝜎 𝜀
l > 3.5(t, w) Along l Thickness shear Thickness shear k15 , d15 , g15 , 𝜖11 , 𝜖11 D
C55 , S55
D
, S55
E
, Q15
mode resonance
𝜎
d > 10t Along t Thickness extension Thickness kt , 𝜖33 D
C33 , C33
E
, S33
E
, Qt
mode resonance
𝜀 𝜎
d > 10t Along t Radial mode Planar resonance kp , 𝜖33 , 𝜖33 𝜎 E , S11
E
, Qp
6.4 Characterization of Piezoelectric Resonator Based on a Resonance Technique 137
Sample Assumption
Poling
t
direction ∼
d 2
D 1
S33 =
𝜌(2tfa )2
D
E
S33 1
S33 = =
2
1 − k33 𝜌(2tfr )2
√
E 𝜎
d33 = k33 S33 𝜖33
d33
g33 = 𝜎
𝜖33
where the dielectric constant 𝜖ij𝜎 can be determined from the capacitance C0𝜎 mea-
sured at 1 kHz, i.e. 𝜖 LF (a frequency substantially lower than the lowest resonance
frequency of the crystal plate), in which case the measurements yield the dielec-
tric permittivity as constant stress. The equation is given by
𝜖ij𝜎 A
C0𝜎 =
t
where A is the area of the electrode. Usually, the constant strain or “clamped”
relative permittivity 𝜖ij𝜀 can be measured at high frequency (frequencies that are
high compared with the principal natural frequencies of the plate but well below
any ionic resonances), i.e. 𝜖HF = 𝜖ij𝜀 .
𝜖ij𝜀 A
C0𝜎 =
t
138 6 Piezoelectric Effect: Basic Theory
𝜀
C0𝜎 t
𝜖HF = 𝜖33 =
A
𝜀
𝜎
𝜖33
𝜖LF = 𝜖33 = 2
1 − k33
E
And its dielectric permittivity and elastic compliance S11 can be determined by
𝜎
C0𝜎 t
𝜖33 = 𝜖LF =
wl
𝜎
𝜖33
𝜖HF = 2
1 − k31
E 1
S11 =
𝜌(2lfr )2
100
1000 000 80
60
40
Impedance Z (Ω)
20
Phase ϕ (º)
100 000
0
–20
–40
10 000
–60
–80
–100
1000
100 200
Frequency f(kHz)
Figure 6.10 Electrical impedance and phase angle versus frequency spectra of a PZT
piezoelectric ceramic rod.
6.4 Characterization of Piezoelectric Resonator Based on a Resonance Technique 139
Poling
direction
Electrodes
w t
∼
l
2
1
d31
g31 = 𝜎
𝜖33
t w ∼
3
2 Poling
direction
140 6 Piezoelectric Effect: Basic Theory
E D 2
C44 = C44 (1 − k15 )
E 1
S44 = E
C44
𝜀
C0𝜎 t
𝜖11 = 𝜖HF =
wl
𝜀
𝜖11
𝜖LF = 2
1 − k15
√
E 𝜎
d15 = k15 S44 𝜖11
d15
g15 = 𝜎
𝜖11
E
C33 = 4𝜌fa2 t 2 (1 − kt2 ) = C33
D
(1 − kt2 )
𝜀
The free relative permittivity 𝜖33 can be calculated from the capacitance C 0 with
the equation given by
𝜀
C0𝜎 t
𝜖33 = 𝜖HF =
A
𝜀
𝜖33
𝜖LF =
1 − kt2
Poling
direction Electrodes
t ∼
d
2
1
References 141
Poling r
direction t ∼
References
Anon (1988). IEEE Standard on Piezoelectricity. In: ANSI/IEEE Std 176-1987.
New York, NY: IEEE Standards Board.
Berlincourt, D.A., Curran, D.R., and Jaffe, H. (1964). Piezoelectric and
piezomagnetic materials and their function in transducers. In: Physical Acoustics:
Principles and Methods (ed. W.P. Mason), 169–270. New York/London: Academic
Press. Available at: https://doi.org/10.1016/B978-1-4832-2857-0.50001-0.
Chen, Y. (2013). High-Frequency and Endoscopic Ultrasonic Transducers Based on
PMN–PT and PIN–PMN–PT Single Crystals. The Hong Kong Polytechnic
University. Available at: http://theses.lib.polyu.edu.hk/handle/200/6985.
142 6 Piezoelectric Effect: Basic Theory
Meitzler, A.H., Obryan, H.M., and Tiersten, H.F. (1973). Definition and
measurement of radial mode-coupling factors in piezoelectric ceramic materials
with large variations in Poisson’s ratio. IEEE Transactions on Sonics and
Ultrasonics 20 (3): 233–239.
Uchino, K. (2009). Ferroelectric Devices, 2e. CRC Press.
143
Piezoelectric Devices
Piezoelectric element
Ultrasonic beam
Input pulse Focusing lens
Figure 7.1 The basic geometry of a single-element ultrasonic transducer while focusing lens
can be convex or concave lens depending on the acoustic velocity between lens and
front-loading medium.
medium allows maximum transmission of the ultrasonic wave, and the backing
layer is to damp the back-radiated acoustic wave to reduce the ring down.
Piezoelectric materials, such as PZT ceramics or relaxor ferroelectric single
crystals, are usually used as the active element to generate and receive ultrasonic
waves. Other piezoelectric materials such as PVDF, LiNbO3 , and lead-free
piezoelectric materials have also been used as active element of ultrasonic
transducers.
Transducers with broad bandwidth (BW) and high sensitivity are important
for ultrasonic imaging applications. Usually one or two acoustic matching layers
are used to increase sound transmission into tissue to compensate the acous-
tic impedance (Z) mismatch between piezoelectric material and human tissue.
Acoustic impedance Z is a parameter used to evaluate the acoustic energy trans-
fer between two materials. Acoustic impedance of a medium is defined as
pressure
Z= (7.1)
volume velocity
For materials,
Z = 𝜌c (7.2)
where 𝜌 and c are the density and the sound velocity, respectively, of the
material. The unit of the acoustic impedance is kg/(m2 s) or Rayl. Table 7.1
lists the acoustic impedance for some nonbiological and biological materials
(Zagzebski 1996).
After applying electrical pulses on the piezoelectric active element, the pres-
sure waves are generated in both back and front directions. The transmission
coefficient (𝛼 T ) evaluates the pressure transmits into the front-loading medium:
4Zpf
𝛼T = (7.3)
(1 + Zpf )2 cos2 (km ⋅ L) + (Zmf + Zpm )2 sin2 (km ⋅ L)
Zi
where Zij = Zj
, and Zp , Zf , and Zm are acoustic impedances of the piezoelec-
tric element, front-loading medium, and matching layer, respectively (refer to
7.1 Piezoelectric Ultrasonic Transducers 145
Air 0.0004
Fat 1.34
Water 1.48
Blood 1.65
Muscle 1.71
Skull bone 7.8
PZT ∼30
PMN-PT ∼28
Epoxy ∼2.0
Figure 7.1). k m denotes the wave numbers in matching layer, and L the thick-
ness of matching layer. The transmission coefficient is a function of the acoustic
impedance, frequency, and the matching layer thickness.
Without the matching layers, the transmission coefficient 𝛼 T is calculated to be
only about 10% for the values Zf ∼ 1.5 MRayl for human tissue and Zp ∼ 30 MRayl
for PZT ceramic element. Therefore, it is necessary to insert a matching layer
between the loading medium and piezoelectric element to improve the trans-
mission coefficient of the ultrasonic wave at the interface (Zhou et al. 2011; Zhu
2008). When the matching thickness is equal to quarter wavelength, i.e. L = 𝜆4 ,
the transmission coefficient of Eq. (7.3) can be simplified to be
4Zpf
𝛼T = (7.4)
(Zmf + Zpm )2
For example, if a matching layer has an acoustic impedance of 15 MRayl with
a quarter wavelength thickness, for the previously mentioned PZT-based trans-
ducer, 𝛼 T can reach more than 55%.
L/2 L/2
+ I1 I2 + Front acoustic
Back acoustic
V1 Z1ʹ v Z2ʹ v V2 port
port
– –
+ I3 C Cʹ I= L
Electrical 0
V3
port
–
1: Φ
+V3–
Figure 7.2 KLM electrical equivalent circuit model for a piezoelectric transducer. Source:
Adapted from Kirk Shung (2015).
This model divides a piezoelectric element into two halves, each is represented
by an acoustic transmission line. The acoustic transmission line serves as a
secondary circuit that is linked with an electrical primary circuit by an ideal
transformer as shown in Figure 7.2 (Kirk Shung 2015).
With the KLM model, the electrical impedance of the transducer is given as
1 1 Z1 Z2
Zin = + + (7.5)
j𝜔C0 j𝜔C ′ 𝛷2 (Z1 + Z2 )
where Z1 and Z2 are the input impedances of the acoustic transmission line look-
ing toward the front acoustic port and back acoustic port, respectively, and C 0
and C′ are the clamped and series capacitance due to the acoustic transmission
line, respectively. 𝛷 is the electromechanical turns ratio.
Desilets et al. also determined the optimum impedance of the matching layer(s)
based on the KLM model (Desilets et al. 1978). For one matching layer, the acous-
tic impedance of the matching material Z m satisfies Eq. (7.6)
In order to enhance front matching, two or more matching layers with quarter
wavelength are often used for transducer fabrication. The acoustic impedances
of two matching layers should be as follows (Desilets et al. 1978):
It should be pointed out that usually two matching layers with quarter wave-
length thickness of each layer are used in commercial product as a tradeoff of
price and performance. More matching layers and even gradient matching with
the impedance ranging from the piezoelectric element to human tissue is more
favorable in performance, but the fabrication process is complicated and cost is
very high.
7.1 Piezoelectric Ultrasonic Transducers 147
Ultrasound analyzer
Ultrasound
absorb tank
Transducer
Frequency (MHz)
5.0 M 10.0 M 15.0 M
1.5
–10
1.0
–20
0.5 –30
Voltage (V)
Amplitude (dB)
–40
0.0
–50
–0.5
–60
–1.0
–70
–1.5 –80
13.0 μ 14.0 μ 15.0 μ 16.0 μ 17.0 μ
Time (s)
(b)
Figure 7.3 (a) Setup of pulse-echo measurement and (b) typical measurement results
showing echo signal and its Fourier transform in frequency domain.
148 7 Piezoelectric Devices
7.1.3.3 Resolution
The axial and lateral resolutions are the distinguishable minimum distance
between two reflectors along the axis and perpendicular to the ultrasound
beam direction, respectively. Axial resolution (in depth) is determined by the
ultrasound pulse duration, while the lateral resolution is related to the beam
width in the lateral direction.
For a circular transducer working at center frequency f c , the axial resolution,
Raxial , and the lateral resolution, Rlateral at the focus point are
c
Raxial = (7.12)
2BW ⋅ fc
𝜆d
Rlateral = (7.13)
a
where c is the sound velocity in the loading medium, BW the bandwidth of the
transducer, 𝜆 the wavelength at the loading medium, d the focal length, and a
is the diameter of the aperture (see Figure 7.4). Therefore, increasing the trans-
ducer’s bandwidth and center frequency will improve the axial and lateral reso-
lutions of the imaging.
Practically, the resolution can be determined by imaging metal wire phantom
and find the resolution as shown in the figure. As shown in Figure 7.4, the resolu-
tion can be determined by measuring the width of the −6 dB points in real space
from the image signal intensity.
7.1 Piezoelectric Ultrasonic Transducers 149
Resolution
6 dB
Rlateral
(a) (b)
Figure 7.5 Different types of medical ultrasound transducers developed from our group. (a)
Single-element focused transducer, (b) single-element rotation transducer for intravascular
imaging, (c) 128-element radial array transducer for endoscopic ultrasound imaging, (d)
annular array for dynamic focusing eye imaging, and (e) an ultrasound image of rabbit eye by
our 40-MHz focus transducer.
7.2.1 Terminologies
There are several terminologies to examine the performance of an USM. Usually,
a larger diameter and larger number of waves result in a larger torque and better
performance of the motor. As indicated, the crests of the waves “push” the rotor
forward for linear or circular motion. It can be predicted that more crests would
provide larger “pushing” force on the rotor.
For an USM, the most important characteristic is the torque – velocity depen-
dence. This is also the most significant factor that describes the performance of an
electromagnetic motor. For USM, another important factor is the admittance Y .
As mentioned, Y can be expressed as G + jB. The admittance helps to understand
the resonance behavior and to find the mechanical quality factor. By plotting the
imaginary part of Y against the real part, a typical Nyquist diagram as shown in
Figure 7.8 is obtained, from which resonance and anti-resonance frequencies,
152 7 Piezoelectric Devices
Pressure
Direction of motion
Electrode
~ A ~ B
A sin ωt B sin ωt
(a)
(b)
Figure 7.7 (a) Schematic diagram of motion between stator and rotor and (b) our fabricated
linear motor based on PZT ceramic generated flexural traveling wave.
Im(Y)
f
fh
j0Cd fs
f1
fp
fa fr
Figure 7.8 Nyquist diagram around its fundamental resonance and anti-resonance
frequencies. Source: Adapted from El and Helbo 2000.
7.2 Ultrasonic Motor 153
blocking capacitance and resistance, and other parameters can be found. The
maximum and the minimum of the motional admittance occur at the mechanical
resonance (series) f s and anti-resonance (parallel) f p , respectively. The resonance
and the anti-resonance of the whole system are given by f r and f a , respectively,
whereas the maximum and the minimum of the total admittance are given by f h
and f l .
The characteristic frequencies that are commonly used in the evaluation of the
equivalent circuit parameters are f s and f p . When the quality factor is sufficiently
high (Qm > 100), the pair (f s , f p ) is obtained from the average of the two other
pairs:
(fh , fl ) + (fr , fa )
(fs , fp ) = (7.14)
2
If the G–B diagram obtained from impedance measurement is larger and more
rounded, the resonant characteristics of the motor is better.
The mechanical quality factor Qm can be determined from either the Nyquist
diagram or the Bode diagram (Figure 7.9). In the Bode diagram, the sharpness of
the admittance around the resonance frequency is used to determine Qm from
the pass band at −3 dB of the frequency response.
fs
Qm = (7.15)
Δf−3 dB
Admittance Y
0 dB
–3 dB
fr Frequency (Hz)
Figure 7.9 Bode diagram around its fundamental resonance and anti-resonance frequencies.
154 7 Piezoelectric Devices
+ 280
– 265
+246
+ 258
– 296
– 280
+ 293
+ 297
(a) – 293 (b)
Case
Rotor Assembly
(c) (d)
Figure 7.10 (a) Diagram of stator and (b) the PZT ring. The numbers indicate d33 measurement
results. (c) the parts and (d) the assembly of the motor.
But in some occasions, to make full use of piezoelectric ceramics, the 34 𝜆 region
would be divided into two sectors and also be poled. In this configuration, the
flexural traveling wave will be produced by exciting both the inputs concurrently.
When an electric field is applied, the positively and negatively poled region
will contract and expand correspondingly. With a 90∘ phase difference between
the two waves, and an unpoled 𝜆/4 region, the flexural waves will travel on
the ceramic surface, and make the teeth-shaped stator vibrate in the same way
driving the rotor to rotate through friction force.
Rs
RL
s
Vin
w Surface acoustic wave
Input IDTs Output IDTs
Figure 7.12 Schematic diagram of Rayleigh wave. Source: Adapted from Gardner et al. (2013).
along the surface and a classical example is the ripples created on the surface of
water by a water droplet. The elastic Rayleigh wave has both a surface-normal
component and a surface-parallel component that is parallel to the direction of
propagation. The energy of the SAW is confined to a zone close to the surface
with a few wavelengths thick.
However, due to the relatively low acoustic velocity in the piezoelectric mate-
rial, the operation frequency is limited in a range of a few hundred MHz. This can-
not satisfy the requirement of high frequency and high bit-rate communication
systems. The increasing demand of high-frequency SAW devices accelerates the
development of SAW devices in the GHz range. The high-frequency SAW devices
may be fabricated by either implementing high-resolution photolithography or by
using high acoustic velocity materials as the substrate with a piezoelectric layer
on top (Nakahata et al. 2003), since the operating frequency is proportional to the
acoustic velocity and inversely proportional to the sum of line width and space
of the IDT (see Figure 7.11). However, the use of high-resolution photolithogra-
phy results in dramatical increase in cost and poorer reliability. Therefore, using
a substrate with high acoustic velocity is preferred, and layer-based SAW devices
have been introduced to achieve this goal.
Most substrates possessing high acoustic velocity are not piezoelectric and
thus must be combined with a piezoelectric layer such as zinc oxide (ZnO), alu-
minum nitride (AlN), lead zirconate titanate (PZT), lithium niobate (LiNbO3 ),
and lithium tantalate (LiTaO3 ) for the generation of the SAW. Consequently, in
the fabrication of high-frequency SAW devices, many combinations of materials
are currently under investigation as potential candidates, such as ZnO/diamond,
LiNbO3 /sapphire, LiNbO3 /diamond, and LiTaO3 /sapphire (Nakahata et al. 2003;
Uchino 2009; Gardner et al. 2013; Campbell 1998a).
In the selection of materials, a systematic study of the required proper-
ties of piezoelectric materials and substrates is needed. Some important
parameters for selecting SAW device structures are (i) acoustic velocity,
(ii) electromechanical coupling coefficient, (iii) film morphology, and (iv)
structural and interfacial qualities. However, none of the material combi-
nations can satisfy all these requirements. The lattice mismatch, thermal
expansion coefficient, and resistance to oxidation of the materials would
directly affect the structure of films and finally adversely affect the per-
formance of the devices. A careful investigation and selection of materials
are essential in the device fabrication. Tables 7.2 and 7.3 compare the rel-
evant properties for some piezoelectric materials and substrates achieved
7.3 Surface Acoustics Wave Devices 157
Shear wave
acoustic velocity
Materials (ms−1 )
Diamond 12 000
Sapphire 6 000
Quartz 3 500
SiC 6 830
from SAW devices, respectively (Nakahata et al. 2003; Gardner et al. 2013;
Campbell 1998b).
lithium niobate consists of a grid of oxygen ion (O2− ) layers in which the lithium
ions (Li+ ) and niobium ions (Nb5+ ) are positioned, as shown in Figure 7.13a.
The displacement of the lithium ions and niobium ions relative to the neutral
configuration (paraelectric phase) gives rise to a dipole moment and thus to the
spontaneous polarization along the c-axis. In the ferroelectric phase as shown
in Figure 7.13b, lithium ions and niobium ions are displaced above or below the
oxygen ion layers. The dipole moment of LiNbO3 is pointing upward so that
the direction of spontaneous polarization is upward. Figure 7.13c shows the
schematic diagram of the crystal structure of lithium niobate in paraelectric
phase where the niobium ions are placed at the center of the oxygen ions
octahedron and the lithium ions are distributed within the next two octahedra.
One possible approach to improve the device performance is to widen the
choice of substrate materials by using piezoelectric films with IDTs to generate
and detect surface waves on non-piezoelectric substrates. LiNbO3 has been
deposited on different substrates or combined with another piezoelectric layer to
enhance the performance of SAW devices. Film and substrate combinations can
be chosen to control the coupling efficiency, to achieve low propagation loss, to
regulate temperature coefficient of delay, and to increase fundamental frequency
and reduce spurious signals. Thus, not many combinations are available. LiNbO3
has also been used as a substrate material incorporating with AlN or ZnO
piezoelectric layer in SAW device application (Wu et al. 2002; Kao et al. 2004;
Kao et al. 2003).
For the LiNbO3 film deposition, it is usually done on the sapphire substrate.
Since these materials are in hexagonal structure and their lattice mismatch are not
so large, high quality films can be obtained by deposition. Shibata et al. claimed
that their grown LiNbO3 films have enough piezoelectricity for SAW device fab-
rication, and it was also shown that different directions of propagation affect
the signal transmission and different modes of acoustic wave transmission were
Li+
Nb5+
P P
O2–
a T < TC T > TC
(a) (b) (c)
Figure 7.13 (a) Stereoscopic view of the ideal crystal stacking of LiNbO3 along the
crystallographic c axis, (b) side view of ferroelectric phase LiNbO3 , and (c) side view of
paraelectric phase LiNbO3 . Source: Kang et al. (2016). Adapted with permission of Royal
Society of Chemistry.
7.3 Surface Acoustics Wave Devices 159
obtained (Shibata et al. 1995). The center frequency of these devices is much
higher than that of the conventional type of SAW devices with similar size of IDT.
S22
Z0 load a1 = 0 Reflected
b2
Reverse
Transmitted S12 Incident
b1 a2
reflected b reflected b
S11 = = 1 S22 = = 2
incident a1 incident a2
a2 = 0 a1 = 0
transmitted b2 transmitted b1
S21 = = S12 = =
incident a1 incident a2
a2 = 0 a1 = 0
b1 = S11 a1 + S12 a2
b2 = S21a1 + S22a2
film on the surface of the LiNbO3 film followed by photolithography and wet
etching process. The designed wavelength of the IDTs was 16 μm (both the line
width and space of IDTs are 4 μm). Figure 7.15 shows the schematic diagram of
LiNbO3 /diamond SAW device.
The performance of the devices is shown in Figure 7.16, where one can see that
the IL is 30 dB at a center frequency of 940 MHz. The calculated phase velocity
(v = 𝜆f ) from the center frequency is 15 000 ms−1 , which are well above the shear
wave sound velocity for diamond and lower than that of the longitudinal wave.
AI IDTs
Δ μm
LiNbO3
piezoelectric film
4 μm
Figure 7.15 Schematic diagram of SAW device structure (LiNbO3 /Al2 O3 /diamond). Inset is the
aluminum IDTs image.
–30
0
–1 –40
S11 (dB)
S21 (dB)
–2
–50
–3
–60
–4
–5 –70
600 800 1000 1200 1400
Frequency (MHz)
Figure 7.16 Scattering parameters S11 response of the SAW device fabricated on LiNbO3 films
deposited diamond substrate. Source: Adapted from Lam et al. (2004).
7.3 Surface Acoustics Wave Devices 161
Figure 7.17 Schematic procedures of the modified dice-and-fill method used for fabricating
1–3 composites based on very fragile piezoelectric materials. Zhou et al. (2012). Reprinted with
permission of Elsevier.
162 7 Piezoelectric Devices
with substitutional doping. Ultimately, more complex solid solutions and doping
schemes will have to be explored because these provide an increasing number
of degrees of freedom for identifying extraordinary properties (Aksel and Jones
2010).
Our group has demonstrated successful application of Pb-free ferroelectric
materials in medical ultrasound transducers. Figure 7.17 shows the diagram of
fabricating NBT-BT/epoxy 1–3 composite for transducer fabrication. This 1–3
composite structure can effectively increase the piezoelectric strain due to the 33
vibration mode of each NBT-BT single crystal element and can also reduce the
acoustic impedance due the low acoustic impedance of epoxy matrix. With such
composite, we have fabricated single-element and array ultrasound transducers,
and the transducers show great performance in both sensitivity and bandwidth
as shown in Figures 7.18 and 7.19.
Matching
layer
NBT-BT / epoxy
1–3 composite
Conductive
adhesive
Backing
material
Coaxial
cable
Figure 7.18 Construction of the NBT–BT/epoxy 1–3 composite single-element and linear array
transducers. Source: Zhou et al. (2012). Reprinted with permission of Elsevier.
0.6
0
0.4
Magnitude (dB)
–10
0.0
–20
–0.2
–0.4
–30
–0.6
9 10 11 12 0 1 2 3 4 5 6 7
Time (µs) Frequency (MHz)
(a) (b)
Figure 7.19 (a) Pulse-echo waveform and (b) frequency spectrum of a single array element of
the fabricated NBT–BT/epoxy 1–3 composite linear array ultrasonic transducer. Source: Zhou
et al. (2012). Reprinted with permission of Elsevier.
References 163
References
Aksel, E. and Jones, J.L. (2010). Advances in lead-free piezoelectric materials for
sensors and actuators. Sensors 10 (3): 1935–1954.
Campbell, C.K. (1998a). Basics of piezoelectricity and acoustic waves. In: Surface
Acoustic Wave Devices for Mobile and Wireless Communications, 19–62. San
Diego: Academic Press.
Campbell, C.K. (1998b). Principles of linear-phase SAW filter design. In: Surface
Acoustic Wave Devices for Mobile and Wireless Communications, 67–96. San
Diego: Academic Press.
Desilets, C.S., Fraser, J.D., and Kino, G.S. (1978). The design of efficient broad-band
piezoelectric transducers. IEEE Transactions on Sonics and Ultrasonics 25 (3):
115–125.
El, G.N. and Helbo, J. (2000). Equivalent circuit modeling of a rotary piezoelectric
motor. IASTED International Conference on Modelling and Simulation
(MS’2000), Pittsburgh, Pennsylvania, USA (May 15–17, 2000).
Gardner, J.W., Varadan, V.K., and Awadelkarim, O.O. (2001). IDT microsensor
parameter measurement. In: Microsensors, MEMS, and Smart Devices, 337–339.
Available at: https://doi.org/10.1002/9780470846087.ch11.
Gardner, J.W., Varadan, V.K., and Awadelkarim, O.O. (2013). Introduction to SAW
devices. In: Microsensors, MEMS, and Smart Devices, 303–317. Wiley-Blackwell.
Available at: https://onlinelibrary.wiley.com/doi/abs/10.1002/9780470846087
.ch9.
Jo, W., Dittmer, R., Acosta, M. et al. (2012). Giant electric-field-induced strains in
lead-free ceramics for actuator applications – status and perspective. Journal of
Electroceramics 29 (1): 71–93. Available at: https://doi.org/10.1007/s10832-012-
9742-3.
Kang, X., Sang, Y., Liu, H. et al. (2016). Formation mechanism and elimination
methods for anti-site defects in LiNbO3 /LiTaO3 crystals. CrystEngComm 18 (42):
8136–8146. Available at: http://dx.doi.org/10.1039/C6CE01306F.
Kao, K.S., Cheng, C.C., Chen, Y.C., and Lee, Y.H. (2003). The characteristics of
surface acoustic waves on AlN/LiNbO3 substrates. Applied Physics A 76 (7):
1125–1127. Available at: https://doi.org/10.1007/s00339-002-2022-3.
Kao, K.S., Cheng, C.C., Chen, Y.C., and Chen, C.H. (2004). The dispersion
properties of surface acoustic wave devices on AlN/LiNbO3 film/substrate
structure. Applied Surface Science 230 (1–4): 334–339. Available at: http://www
.sciencedirect.com/science/article/pii/S0169433204001850.
Kirk Shung, K. (2015). Diagnostic Ultrasound: Imaging and Blood Flow
Measurements, 2e. Boca Raton, London, New York: CRC Press.
Kosec, M., Malic, B., Wolny, W. et al. (1998). Effect of chemically aggressive
environment on the electromechanical behaviour of modified lead titanate
ceramics. Journal of the Korean Physical Society 32: S1163–S1166.
Krimholtz, R., Leedom, D.A., and Matthaei, G.L. (1970). New equivalent circuits for
elementary piezoelectric transducers. Electronics Letters 6 (13): 398–399.
Lam, H.K., Dai, J.Y., and Chan, H.L.-W. (2004). Highly-oriented LiNbO3 films on
polycrystalline diamond substrate for high frequency surface acoustic wave
devices. Japanese Journal of Applied Physics 43 (6A): L706–L708. Available at:
http://dx.doi.org/10.1143/JJAP.43.L706.
164 7 Piezoelectric Devices
Morgan, D.R. (1973). Surface acoustic wave devices and applications. Ultrasonics 11
(3): 121. Available at: http://www.sciencedirect.com/science/article/pii/
0041624X73906082.
Nakahata, H., Fujii, S., Higaki, K. et al. (2003). Diamond-based surface acoustic wave
devices. Semiconductor Science and Technology 18 (3): S96–S104. Available at:
http://stacks.iop.org/0268-1242/18/i=3/a=314.
Reilly, N.H.C., Milsom, R.F., and Redwood, M. (1973). Generation of Rayleigh and
bulk waves by interdigital transducers on y-cut z-propagating lithium niobate.
Electronics Letters 9 (18): 419–420.
Rödel, J., Jo, W., Seifert, K.T.P. et al. (2009). Perspective on the development of
lead-free piezoceramics. Journal of the American Ceramic Society 92 (6):
1153–1177. Available at: http://dx.doi.org/10.1111/j.1551-2916.2009.03061.x.
Rödel, J., Webber, K.G., Dittmer, R. et al. (2015). Transferring lead-free piezoelectric
ceramics into application. Journal of the European Ceramic Society 35 (6):
1659–1681. Available at: http://www.sciencedirect.com/science/article/pii/
S0955221914006700.
Saito, Y., Takao, H., Tani, T. et al. (2004). Lead-free piezoceramics. Nature 432: 84.
Available at: http://dx.doi.org/10.1038/nature03028.
Sashida, T. and Kenjo, T. (1991). An Introduction to Ultrasonic Motors. Clarendon
Press.
Shibata, Y., Kaya, K., Akashi, K. et al. (1995). Epitaxial growth and surface acoustic
wave properties of lithium niobate films grown by pulsed laser deposition.
Journal of Applied Physics 77 (4): 1498–1503. Available at: https://doi.org/10
.1063/1.358900.
Uchino, K. (2009). Ferroelectric Devices, 2e. CRC Press.
Wong, K.-K. (2002). Acoustic wave propagation and properties. In: Properties of
Lithium Niobate. London: INSPEC, The Institution of Electrical Engineers.
Wu, S., Chen, Y.-C., and Chang, Y.S. (2002). Characterization of AlN films on Y-128∘
LiNbO3 by surface acoustic wave measurement. Japanese Journal of Applied
Physics 41 (Part 1, No. 7A): 4605–4608. Available at: http://dx.doi.org/10.1143/
JJAP.41.4605.
Zagzebski, J.A. (1996). Essentials of Ultrasound Physics. USA: Mosby: Maryland
Heights. Available at: https://books.google.com.hk/books?id=XahsQgAACAAJ.
Zhou, Q., Lam, K.H., Dai, J.Y. et al. (2011). Piezoelectric films for high frequency
ultrasonic transducers in biomedical applications. Progress in Materials Science
56 (2): 139–174. Available at: http://www.sciencedirect.com/science/article/pii/
S0079642510000563.
Zhou, D., Lau, S., Wu, D., and Kirk Shung, K. (2012). Lead-free piezoelectric single
crystal based 1–3 composites for ultrasonic transducer applications. Sensors and
Actuators A: Physical 182: 95–100. Available at: http://www.sciencedirect.com/
science/article/pii/S0924424712003299.
Zhu, J. (2008). Optimization of Matching Layer Design for Medical Ultrasonic
Transducer. The Pennsylvania State University.
Zhu, B.P., Wu, D.W., Zhou, Q.F. et al. (2008). Lead zirconate titanate thick film with
enhanced electrical properties for high frequency transducer applications.
Applied Physics Letters 93: 012905. Available at: https://doi.org/10.1063/1
.2956408.
165
Magnetic Magnetic Ms
moment m moment m Mr
Electron
–Hc +Hc H
Atomic Direction
nucleus of spin
Figure 8.1 Diagram showing the magnetic moment associated with (a) orbital motion and (b)
spin motion of an electron; (c) ferromagnetic M–H loop where Ms and Mr are the saturation
and remnant magnetization and Hc is the coercive magnetic field.
m m
Ferromagnetism Ferroelectricity
they similarly have properties of being switchable and vanished above Curie tem-
perature. There are phenomenological similarities between ferromagnetic and
ferroelectric, where ferromagnetic is due to time-reversal symmetry breaking and
ferroelectric originates from spatial inversion symmetry breaking. More analo-
gies can be seen from Table 8.1 comparing their fundamental properties.
Electrons have an intrinsic property called spin that contributes to their mag-
netic moment. The classical model of electron spin is the electron spinning on its
axis, i.e. either spin-up or spin-down. Then the value of magnetic moment can be
calculated with respect to the spin of an electron. The total magnetic moment of
an atom is the vector sum of the orbital and spin magnetic moments. A magnetic
solid, made up of atoms with magnetic moment, has quantum exchange interac-
tions that tend to align the magnetic moments at low temperature. When T < T c ,
the macroscopic magnetization arises and retains even in the absence of a mag-
netic field. The magnetic moments tend to align in the same direction without the
aid of an external magnetic field. This is known as the ferromagnetic phase. In the
ferromagnetic category, materials are divided into strong and weak ferromagnets.
By introducing the total particle number N = N ↑ + N ↓ and the spin polarization
s = N ↑ − N ↓ , a strong ferromagnet has almost s = 100% at the Fermi energy as
shown in Figure 8.3a, while a weak ferromagnet gets a smaller spin polarization
and paramagnet has no net spin polarization (Figure 8.3b,c).
Meanwhile, there is another type of magnets that possess magnetization
microscopically, but have no macroscopic magnetization due to the cancelation
of antiparallel spins of neighboring pairs – this is known as AFM phase. A
magnet that exhibits no macroscopic magnetization at high temperature (when
H = 0) is called as the paramagnetic phase, in which the magnetic moment
induced by the applied H-field is rather weak. Figure 8.4 shows the relationship
between magnetization and temperature of a ferromagnet, where one can see
that the magnetization starts decreasing close to T c , and when T > T c , magnetic
168 8 Ferromagnetics: From Material to Device
E E E
EF EF
Figure 8.3 Schematic densities of states (DOSs) for ferromagnetism: (a) strong ferromagnet,
(b) weak ferromagnet, and (c) paramagnet.
Mr
Tc T
Figure 8.4 Illustration of magnetization versus temperature. Arrows inside the boxes denote
the magnetic moments alignment.
T > Tc
T = Tc
Energy
T < Tc
Magnetization
Figure 8.5 Sketch of Landau free energy F(M) = 𝛼M2 + 𝛽M4 for T > T c , T = T c , and T < Tc .
order parameter:
F(M) = F0 + 𝛼M2 + 𝛽M4 + … (8.1)
F is minimized with respect to M:
𝜕F
= 2𝛼M + 4𝛽M3 = 0 (8.2)
𝜕M
𝛼
M = 0 or M2 = − (8.3)
2𝛽
In Figure 8.5, the symmetry changes precisely at T = T c and the transition occurs
when 𝛼 changes its sign. When T < T c , there are two energy minima M = ± Ms
that indicate 𝛼 < 0 and 𝛽 > 0. When T > T c , a single energy minimum M = 0
indicates 𝛼 > 0 and 𝛽 > 0. The temperature dependence of 𝛼 can be identified as
( )
T − Tc
𝛼(T) = 𝛼0
Tc
With the presence of an external magnetic field H, a linear term is added as
F(M) = 𝛼M2 + 𝛽M4 + … − 𝜇0 HM (8.4)
As shown in Figure 8.6, when an external magnetic field breaks the symmetry,
the global minimum of F(M) changes from two minima to a single minimum.
Therefore, we can see that even when T > T c , a paramagnetic phase of material,
which has no macroscopic magnetization, can also be attracted with the presence
of an externally applied magnetic field.
T > Tc
T = Tc
Energy
T < Tc
Magnetization
Figure 8.6 Sketch of Landau free energy F(M) = 𝛼M2 + 𝛽M4 − 𝜇0 HM for T > T c , T = T c , and
T < T c under a positive magnetic field.
There are easy axes in ferromagnetic materials that are the directions magnetic
moments should follow. For example, for a cubic structured FM material such as
Fe, the ⟨100⟩ directions are usually the easy axes, while for a hexagonal-structured
FM material such as Co, the ⟨0001⟩ are the easy axes. Due to this difference, in a
cubic structure, both 90∘ and 180∘ magnetic domains can be formed (Figure 8.7a),
while in a hexagonal structured ferromagnetic material, the domains are usually
aligned with an angle of 180∘ (Figure 8.7b).
The easy axes of cubic-structured Ni are ⟨111⟩, so it can form 180∘ , 71∘ , and
109∘ magnetic domains; these are possible angles between all ⟨111⟩ easy axes. The
magnetization switching can be realized by not only external magnetic field but
also mechanical stress. This mechanical stress-induced magnetization switching
is similar to ferroelectric material whose polarization can be switched by both
electric field and mechanical stress.
(a) (b)
Figure 8.7 Domain structures in cubic (a) and hexagonal (b) structured ferromagnetic
materials.
8.4 Magnetoresistance Effect and Device 171
Figure 8.8 Domain wall structures of (a) Bloch-type ferromagnetic domain walls, (b) Néel-type
ferromagnetic domain walls, and (c) Ising-type ferroelectric domain walls.
H-field direction
H-field
M
direction
M
e–
e–
θ
(a) (b)
Figure 8.9 Schematic diagram of origin of the AMR effect, (a) high and (b) low resistance
states raised by the current parallel and perpendicular to the magnetization, respectively.
Ovals represent the scattering cross-section of the bound electronic orbits.
a material rotates, the electron cloud of each nucleus deforms and it changes
the amount of conduction electron scattering, i.e. the magnetization direction
rotates the closed orbit orientation with respect to the current direction. As
demonstrated in Figure 8.9a, if the field and magnetization are oriented parallel
to the current, the electronic orbits are oriented perpendicular to the current,
so the cross-section for scattering increases, resulting in a high resistance state.
On the contrary, if the field is transverse to the current, the electronic orbits
are in the plane with current; therefore, less scattering occurs due to the small
cross-section, resulting in a low resistance state (Figure 8.9b).
The relation of electrical resistance R on the angle 𝜃 between the directions of
electrical current and magnetization can be obtained by
R = R⟂ + (R∥ − R⟂ )cos2 𝜃 (8.6)
Owing to the angular dependence, AMR effect is popularly used in
magneto-resistor and sensors. For example, electronic compass for the
measurement of Earth’s magnetic field, traffic detection, and linear position
and angle sensing. A typical AMR magnetoresistive sensor is constructed of a
combination of magneto-resistors, which can be used to measure strength and
direction of magnetic field.
Besides sensors, AMR effect was used as the basis of reading in magnetore-
sistive random-access memory (MRAM). However, its amplitude in ferromag-
netic materials is weak (about 2% variations in resistance on changing the rela-
tive orientation of magnetization and current). The application in reading head
using AMR effect was therefore replaced after the GMR was discovered. How-
ever, AMR effect has attracted great interest again because of investigation on
AFM materials (Kriegner et al. 2016; Lu et al. 2018; Hu et al. 2012). It represents
a major step in the field of AFM spintronics in which the antiferromagnet (with
a zero net moment) governs the transport instead of just playing a passive sup-
porting role in the traditional ferromagnetic (FM) spintronics.
conductive layers (Baibich et al. 1988). Unlike AMR, GMR requires nano-
structured composites and each layer is only a few nanometers thick. The
magnetoresistance was reported as the largest so far observed for magnetic
metal films (about 40% for Fe/Cr multilayers), so it was called as GMR. The 2007
Nobel Prize in Physics was awarded to Albert Fert and Peter Grünberg for the
discovery of GMR.
Let us look at the details of the work being awarded 2007 Nobel Prize in Physics.
As reported in 1987–1988 (Grünberg et al. 1987; Baibich et al. 1988), they grew
Cr/Fe superlattice structure and measured the in-plane resistance in the presence
of magnetic field. Due to the exchange coupling between the FM Fe layer and Cr
layer, this superlattice structure can have parallel and antiparallel magnetization
layers forming a GMR effect as illustrated in Figure 8.10. One can see that the
maximum magnetoresistance reaches 80% from the superlattice structure with
Cr layer thickness of 0.5 nm. This GMR effect has been successfully implemented
in hard disc drive date storage.
Not only ferromagnetic metals, GMR was also found to occur in some
oxide thin films of manganates of the formula La1−x Ax MnO3 (A = alkaline
earth metal). One of the well-known examples is La1−x Srx MnO3 . The parent
manganate LaMnO3 , AFM insulator, is an orthorhombic perovskite with dis-
torted MnO6 octahedra. When La is replaced by a divalent ion, Mn4+ ions are
created, and when Mn4+ is sufficiently large, the material changes from AFM
to ferromagnetic, exhibiting an insulator-to-metal transition because of double
exchange mechanism. The magnetization and electrical resistivity behavior of
R/R(H = 0)
1.0
0
0.8
0 .8 ~+80%
8
3 nm)/Cr(1.2 nm)
Fe(3 0.7
0.6
(0.9 nm)
Fe(3 nm)/Cr(0.9
0.5
–40 –20 0 20 40
H (kG)
Figure 8.10 Magnetoresistance of Fe/Cr superlattice showing the change in the resistance of
Fe/Cr superlattice at 4.2 K in external magnetic field H. The arrow indicates maximum
resistance change. Source: Adapted from Baibich et al (1988).
174 8 Ferromagnetics: From Material to Device
La1−x Ax MnO3 offers the oxide thin films an opportunity for ferromagnetism and
GMR effect applications (Rao and Mahesh 1997).
The nature of the GMR effect is different from that of AMR – that the
resistance in the plane of the film is isotropic for the GMR effect but depends
on current direction for the AMR. The GMR effect shows a magnetization
orientation-dependent component of resistance that varies between magnetiza-
tions in alternate layers. Figure 8.11a shows the schematic superlattice structure
of ferromagnetic Fe layers alternatively with Cr layers reported by Baibich et al.
In this structure, layers of Fe, sandwiching an ultrathin Cr layer, was found to be
ferromagnetically ordered with a magnetic moment oriental antiparallel to that
Fe
Cr
Fe
Cr
Fe
Cr
Fe
GaAs
When H ≠ 0, When H = 0,
Fe Fe
Cr Cr
Fe Fe
Rparallel Rantiparallel
(b) (c)
Figure 8.11 (a) Diagram showing the superlattice structure of ferromagnetic Fe layers
alternatively with Cr layers, white arrows denote magnetization direction, and (b) and (c) are
the corresponding low and high resistance states for parallel and antiparallel magnetizations
among neighboring Fe layers.
8.4 Magnetoresistance Effect and Device 175
GMR sensor
AC generator
Figure 8.12 Hybrid probe: eddy current testing (ECT) coil with GMR sensor. Source: Adapted
from Čápová et al. (2008).
of the adjacent Fe layer. Without applied magnetic field, the coupling of Fe films
across Cr interlayers with proper thickness is found to be AFM (Saurenbach
et al. 1988; Grünberg et al. 1987), and it brings the magnetization to zero. GMR
effect is thus achieved in antiferromagnetically coupled Fe/Cr superlattice by
aligning the magnetization of adjacent Fe layers with an external field.
By well controlling the AFM coupling, the ferromagnetic layers can form a state
with antiparallel magnetization. Electrons that transport through this structure
will be highly scattered due to the averaged spin-up and spin-down currents,
resulting in high resistance state (Figure 8.11c). When a magnetic field is applied
in parallel to the film, AFM magnetization can be switched to the magnetic field
direction, making the whole film’s magnetization direction into the field direction
as shown in Figure 8.11b; electrons transport through this layer experience weak
interface scattering and a high transmission for the spin-down electrons, result-
ing in a relatively much lower resistance. When the thickness of the Cr layer or the
temperature increases, more defect and steps may be introduced, which results in
weakened AFM coupling within the Cr layer; this makes the GMR effect relatively
smaller. This model explains the GMR and explains the thickness-dependence of
GMR shown in the diagram.
A GMR sensor has large bandwidth making it applicable to different frequen-
cies of magnetic field. GMR sensors also have high sensitivity and low power
consumption. Figure 8.12 shows an eddy current probe and a GMR magnetic
field sensor being used in power line current stability monitoring by eddy current
detection. GMR sensors have also been used in smart parking system where the
approaching and leaving of vehicles can be sensed by measuring the MR due to
the Earth’s magnetic field change induced by the vehicles’ movement.
E E E E
eU
eU
DOS of pinned layer DOS of free layer DOS of pinned layer DOS of free layer
Figure 8.13 Diagram of spin-dependent TMR effect in parallel (left) and antiparallel (right)
magnetization of a MTJ. The solid arrow represents larger tunneling probability.
Bit line
Gate Gate
(a) (b)
Figure 8.14 Diagram of (a) a conventional MRAM cell and (b) STT-RAM cell memory structures.
and only happens below Curie temperature when a magnetic material is in the
ferromagnetic phase. This effect exists in all magnetic materials either smaller or
larger and can be positive and negative. The mechanical strain induced by mag-
netic field is called direct magnetostrictive effect, while the magnetization change
resulted from applied mechanical strain is called inverse magnetostrictive effect.
You may have experience to hear the low-frequency noise “electric hum”
near transformers and high-power electrical devices. That sound comes from
the magnetostrictive vibration of magnetic core material in the transformer
at a frequency of power line that is usually 50 Hz. For device applications, in
order to have large response to a magnetic field making it applicable for sensors
or actuators, the magnetostrictive effect should be significant. Table 8.2 lists
properties of magnetostrictive effect of some typical magnetic materials.
In this table, 𝜀max = 32 𝜆s is the maximum strain that can be generated under a
prestress condition, where 𝜆s is the saturation magnetostriction that quantifies
strains generated from a relaxed ferromagnetic state to the state with all domains
aligned along the magnetic field direction. Ms is the saturation magnetization in
response to an applied magnetic field, which should be large for strong transduc-
tion; and T c is the Curie temperature at which a ferromagnetic to paramagnetic
phase transition occurs. The magnetomechanical coupling coefficient k quanti-
fies the efficiency to convert magnetic energy to mechanical energy, or vice versa,
and elastic modulus Y is its mechanical properties quantifying the strain that can
be generated under mechanical stress. These parameters are important for mag-
netostrictive material-based transducer design.
This table provides general guidelines to select a suitable magnetostrictive (MS)
material for device application. Fe, Ni, Co, and CoFe2 O4 possess negative and
much smaller value of maximum strain and usually small magnetomechanical
coupling factor k. These parameters make them less attractive for device applica-
tion. Tb and Dy have very large magnetostriction, but their T c is too low to make
them meaningful for device application.
The most significant one is Terfenol-D (terbium: Ter, iron: Fe, Naval Ordnance
Laboratory: NOL, dysprosium: D), which can produce strains up to 1600 μl/l at
room temperature and therefore can greatly expand the role of magnetostrictive
smart material for transducer applications. Metglas is also practically very popu-
lar since it has the largest magnetomechanical coupling factor k, making it more
efficient and having broader bandwidth of response to magnetic field.
[112]
[111]
[001]
[112]
Prestress
[110]
[111] 2a
a
Figure 8.15 Terfenol-D crystal orientation where [112] is the rod direction and ⟨111⟩ are
magnetization easy axes.
8.5 Magnetostrictive Effect and Device Applications 181
λs = 2 ε
3
𝜀(𝜃) = 𝜀 cos2 𝜃
This equation is only suitable for the case when an isotropic paramagnetic (repre-
sented by the round circle in the diagram) changes to isotropic FM (represented
by ellipsoids). Without external magnetic field, these domains, each of them is
represented by one ellipsoid, are randomly orientated, and the spontaneous strain
182 8 Ferromagnetics: From Material to Device
where one can see that the spontaneous strain is 1/3 of the total strain. When
these domains align their magnetization along external magnetic field, the extra
strain (called saturated strain 𝜆s ), is then equal to 2/3 of the total strain, i.e.
2
𝜆s = 𝜀 − 𝜆 0 = 𝜀
3
Figure 8.16d illustrates the maximum strain that can be achieved from such
isotropic magnetostrictive material through prestress to force all domains
into alignment to the direction perpendicular to the external magnetic field,
subsequently applied to rotate the magnetization moment by 90∘ .
As illustrated in Figure 8.16, it also suggests the practical way to measure the
total strain by measuring the difference of strains when the magnetic field is,
respectively, perpendicular and parallel to the rod, i.e.
3
𝜆s|| − 𝜆s⟂ = 𝜆 =𝜀
2 s
The state 𝜆s⟂ can also be obtained by applying prestress forcing the domains into
a state with magnetic moment along the perpendicular direction as illustrated by
the ellipsoids in Figure 8.16d. One can imagine that when the parallel magnetic
field and the prestress along the rod direction reach a metastable sate, minimum
perturbation of ac magnetic field will cause the magnetic momentum to rotate
from perpendicular to parallel or from parallel to perpendicular. In fact, this is
the way to achieve maximum strain in real application that will be further demon-
strated in next chapter.
Another important relationship can be derived from
𝜀(θ) = 𝜀 cos2 𝜃
By substituting with
M = Ms cos 𝜃
and then,
3
𝜆(𝜃) = 𝜆 cos2 𝜃
2 s
This results in a quadratic relationship between the magnetostriction and the
𝜆
magnetization, i.e. 𝜆(M) = 32 Ms M2 .
s
One should notice that electrostrictive effect is also the quadratic relation.
Compared with P–E loop and 𝜀 − E butterfly loop of ferroelectric materials,
M–H and 𝜀 – H exhibit similar hysteresis and butterfly loops as shown in the
Figure 8.17.
8.5 Magnetostrictive Effect and Device Applications 183
1.0
0.8
0.6
0.4
0.2
M/Ms
–0.2
–0.4
f = 1 Hz
–0.6 f = 500 Hz
f = 1000 Hz
–0.8 f = 2000 Hz
–1.0
–1.5 –1 –0.5 0 0.5 1 1.5
(a) H (A/m) ×105
1800
f = 1 Hz
1600
f = 500 Hz
f = 1000 Hz
Magnetostrictive strain (10–6)
1400
f = 2000 Hz
1200
1000
800
600
400
200
0
–1.5 –1 –0.5 0 0.5 1 1.5
(b) H (A/m) ×105
Figure 8.17 The magnetic (a) hysteresis loops and (b) magnetostrictive strain hysteresis loops
of Terfenol-D under the sinusoidal driving magnetic field of same amplitude and different
frequencies. Source: Xiao et al. (2017). Reprinted with permission of Elsevier.
Solenoid with
electrical ac input
Magnetostrictive
laminations
Magnetization
and strain induced
by magnetic field
Force
Metglas ribbon
V
Detection coil
Excitation coil
shows a diagram of the force sensors using magnetoelastic effect. When a force
is added, the magnetoelastic material deforms, and as a result, the strain can be
detected by measuring the flux density change.
Sample
Pick-up coils
200
50
0 0
–50
–200
–100
–400
–150
–200 0 200 –200 0 200
Magnetic field (Oe) Magnetic field (Oe)
(b) (c)
Figure 8.20 (a) Schematic diagram of VSM system and M–H loops of two magnetic materials
LSMO and NiFe in in-plane (b) and (c) out-of-plane magnetization directions at 78 K.
the out-of-plane measurement. Based on the results, VSM can also be used for
specifying the easy axis of magnetization for materials. Using inductive method,
for instance, VSM, FM properties including the direction of easy axis, satura-
tion magnetization, and coercivity can be obtained by analyzing M–H loops of
materials. Other common inductive methods include extraction magnetometry,
AC susceptometry, and superconducting quantum interference device (SQUID)
magnetometry.
Josephson junctions
Ia
Is
I I
Φ
Ib Is
Superconducting loop
25
20
Path of cantilever
μm
15
10
0
Magnetic sample with domains 0 5 10 15 20 25 30
(a) (b) μm
Figure 8.22 (a) Schematic diagram of MFM mechanism and (b) MFM image of a magnetic
recording disc displaying a multi-scale domain structure.
Magnetic sample
with domains
(c)
(b)
(a)
Figure 8.23 Schematic illustration showing the optical experiment for observing the (a) polar,
(b) longitudinal, and (c) transverse Kerr effects.
using conventional imaging optics or a finely focused laser spot across the
sample surface. Therefore, the technique is moderately surface sensitive and can
be used to image magnetic domains in thin films that are only a few monolayers
thick.
B M
–F –
– – F
– –
– – –
– –
E –
t w E
t w
B=0
I I
(a) (b)
Figure 8.24 (a) Ordinary Hall effect and (b) anomalous Hall effect (AHE). B is the magnetic
induction rise by external magnetic field; M is the intrinsic spontaneous magnetization.
where the first term represents the ordinary Hall effect mentioned, while the sec-
ond term represents the AHE contribution due to the spontaneous magnetization
as illustrated in Figure 8.24b. Unlike R0 that only depends on the carrier density,
R1 depends on the effects of the spin–orbit interaction of spin electrons (Karplus
and Luttinger 1954). As a result, specific magnetic parameters of FM materials,
such as magnetization processes and types of materials, can be determined using
AHE technique.
σ
Spin current Is Electromotive
σ force EISHE
+
–
Spin curr
ent I
s
(b)
Figure 8.25 (a) Spin Hall effect and (b) inverse spin Hall effect, 𝝈 denotes the spin polarization
vector of the spin current.
References 193
colliding with the impurities within the material. The second possible mecha-
nism is caused by the intrinsic properties of the material that the trajectories
of carriers are distorted due to spin–orbit interaction as a consequence of the
asymmetries in the material. Owing to the accumulation of spins of both lateral
boundaries, the current and spin Hall voltage can be detected. By the inverse
spin Hall effect (ISHE, Figure 8.25b) in a strong spin–orbit coupling material,
the spin–orbit interaction bends trajectories of these two electrons in the same
direction. The spin current can then be converted into an electric field EISHE with a
charge current again in the transverse direction, resulting in charge accumulation
at the sample edges.
For the SHE, no magnetic field is required as it is a purely spin-based phe-
nomenon dissimilar to the ordinary Hall effect, but it behaves similar to AHE,
known for a long time in ferromagnet, which also originates from spin–orbit
interaction. SHE can be used to manipulate electron spins electrically, and a pure
spin current can even be generated by non-magnetic materials.
References
Baibich, M.N., Broto, J.M., Fert, A. et al. (1988). Giant magnetoresistance of
(001)Fe/(001)Cr magnetic superlattices. Physical Review Letters 61 (21):
2472–2475.
Čápová, K., Čáp, I., Janoušek, L., and Smetana, M. (2008). Recent trends in
electromagnetic non-destructive sensing. Advances in Electrical and Electronic
Engineering 7 (1): 322–325. Available at: http://hdl.handle.net/10084/84200.
Celotta, R.J., Unguris, J., Kelley, M.H., and Pierce, D.T. (2002). Characterization of
Materials. In: Techniques to measure magnetic domain structures (ed. E.N.
Kaufmann), 1. https://doi.org/10.1002/0471266965.com046. Available at: http://
doi.wiley.com/10.1002/0471266965.com046.
Daughton, J.M. (1992). Magnetoresistive memory technology. Thin Solid Films 216:
162–168.
Eerenstein, W., Mathur, N.D., and Scott, J.F. (2006). Multiferroic and
magnetoelectric materials. Nature 442 (7104): 759–765. Available at: http://www
.ncbi.nlm.nih.gov/pubmed/16915279.
García-Martín, J., Gómez-Gil, J., and Vázquez-Sánchez, E. (2011). Non-destructive
techniques based on eddy current testing. Sensors 11 (3): 2525–2565. Available at:
http://www.mdpi.com/1424-8220/11/3/2525.
Grünberg, P., Schreiber, R., Pang, Y. et al. (1987). Layered magnetic structures:
evidence for antiferromagnetic coupling of Fe layers across Cr interlayers. Journal
of Applied Physics 61 (8): 3750–3752.
Hu, J., Li, Z., Chen, L., and Nan, C. (2012). Design of a voltage-controlled magnetic
random access memory based on anisotropic magnetoresistance in a single
magnetic layer. Advanced Materials 24 (21): 2869–2873. Available at: https://
onlinelibrary.wiley.com/doi/abs/10.1002/adma.201201004.
Inoue, J.I. (2013). GMR, TMR, BMR, and related phenomena. In: Nanomagnetism
and Spintronics, 2e, 15–106. Elsevier B.V. Available at: http://dx.doi.org/10.1016/
B978-0-444-63279-1.00002-2.
194 8 Ferromagnetics: From Material to Device
Jiles, D. (2015). Introduction to Magnetism and Magnetic Materials, 3e. CRC Press.
Jiles, D.C. (1995). Theory of the magnetomechanical effect. Journal of Physics D:
Applied Physics 28: 1537–1546.
Jin, S., Tiefel, T.H., McCormack, M. et al. (1994). Thousandfold change in resistivity
in magnetoresistive La-Ca-Mn-O films. Science 264 (5157): 413–415. Available at:
http://www.sciencemag.org/cgi/doi/10.1126/science.264.5157.413.
Julliere, M. (1975). Tunneling between ferromagnetic films. Physics Letters A 54 (3):
225–226.
Karplus, R. and Luttinger, J.M. (1954). Hall effect in ferromagnetics. Physical Review
95 (5): 1154–1160.
Kiselev, S.V., Ozerov, R.P., and Zhdanov, G.S. (1963). Detection of magnetic order in
ferroelectric BiFeO3 by neutron diffraction. Soviet Physics – Doklady 7 (8):
742–744.
Kriegner, D., Výborný, K., Olejník, K. et al. (2016). Multiple-stable anisotropic
magnetoresistance memory in antiferromagnetic MnTe. Nature Communications
7: 11623. Available at: http://dx.doi.org/10.1038/ncomms11623.
Liu, Y., Tang, N., Wan, X. et al. (2013). Realization of ferromagnetic graphene oxide
with high magnetization by doping graphene oxide with nitrogen. Scientific
Reports 3: 2566. Available at: http://dx.doi.org/10.1038/srep02566.
Lu, C., Gao, B., Dong, S. et al. (2018). Revealing controllable anisotropic
magnetoresistance in spin–orbit coupled antiferromagnet Sr2 IrO4 . Advanced
Functional Materials 28 (17): 1706589. Available at: https://onlinelibrary.wiley
.com/doi/abs/10.1002/adfm.201706589.
Ma, Y.W., Lu, Y.H., Ding, J. et al. (2012). Room temperature ferromagnetism in
Teflon due to carbon dangling bonds. Nature Communications 3: 727. Available
at: http://dx.doi.org/10.1038/ncomms1689.
Mazin, I.I. (1999). How to define and calculate the degree of spin polarization in
ferromagnets. Physical Review Letters 83 (7): 1427–1430.
Nickel, J. (1995). Magnetoresistance Overview: HP Laboratories Technical Report:
HPL-95-60.
Poon, T.Y., Tse, N.C.F., and Lau, R.W.H. (2013). Extending the GMR current
measurement range with a counteracting magnetic field. Sensors 13 (6):
8042–8059. Available at: http://www.mdpi.com/1424-8220/13/6/8042.
Pugh, E.M. and Rostoker, N. (1953). Hall effect in ferromagnetic materials. Reviews
of Modern Physics 25 (1): 151.
Rao, C.N.R. and Mahesh, R. (1997). Giant magnetoresistance in manganese oxides.
Current Opinion in Solid State and Materials Science 2 (1): 32–39. Available at:
http://www.sciencedirect.com/science/article/pii/S1359028697801027.
Saurenbach, F., Walz, U., Hinchey, L. et al. (1988). Static and dynamic magnetic
properties of Fe-Cr-layered structures with antiferromagnetic interlayer
exchange. Journal of Applied Physics 63 (8): 3473–3475.
Savage, H.T. and Spano, M.L. (1982). Theory and application of highly
magnetoelastic Metglas 2605SC. Journal of Applied Physics 53: 8092.
Spinu, L., Dodrill, B.C., and Radu, C. (2013). Magnetometry measurements. In: ,
1–3. Magnetics Technology International.
Tsai, C.Y., Chen, H.R., Hsieh, W.F. et al. (2013). Stress-mediated magnetic
anisotropy and magnetoelastic coupling in epitaxial multiferroic
References 195
P M
H E
Figure 9.1 Illustration of multiferroics with expected P–H and M–E hysteresis loops.
under space and time inversions. Space inversion will reverse the direction of
polarization P, while leaving the magnetization M invariant. Time reversal,
in turn, will change the sign of M, while the sign of P remains invariant.
Magnetoelectric multiferroics require simultaneous violation of space and
time inversion symmetries. In BiFeO3 , for example, off-centering of ions gives
rise to an electric polarization, while additional magnetic ordering breaks
time-reversal symmetry. Therefore, BiFeO3 has been extensively studied as
a model multiferroic material that will be particularly introduced in the
following text.
Although scientists have already started experimental observation and theoret-
ical prediction of multiferroics in single phase materials many decades ago, the
term multiferroics was firstly used by H. Schmid in 1994 (Schmid 1994). Since
then, the concept of multiferroics has been expanded to materials that perform
any type of long-range magnetic ordering, a spontaneous electric polarization,
and/or ferroelasticity, and some researchers argue that the term multiferroics
can only be used in materials that exhibit coupling of the ferroic order parame-
ters. The overlap of these properties in ferroic materials is illustrated in Figure 9.2
(Eerenstein et al. 2006).
Multiferroic
Ferroelectric Ferromagnetism
Electrically Magnetically
polarizable polarizable
Magnetoelectric
Figure 9.2 Relationship between ferroelectrics, ferromagnetics, multiferroics, and even the
magnetoelectrics. Source: Adapted from (Eerenstein et al. (2006)).
9.3 Why Are There so Few Magnetic Ferroelectrics? 199
120
BFO(111)
BFO(110)
80 BFO(100)
Polarization (μC/cm2)
40
–40
–80
–120
–10 –5 0 5 10
Applied voltage (V)
Figure 9.3 Ferroelectric hysteresis loops of epitaxial BFO thin films along different crystalline
directions. Source: Chu et al. (2007). Reprinted with permission of Elsevier.
Pi+
(b)
Bi
Fe
(a) O
(a)
(b)
(c)
Figure 9.5 Domain structures in BFO with (a) 71∘ , (b) 109∘ , and (c) 180∘ domain walls. Source:
Adapted from Baek et al. (2010).
[111]
MFe1
MFe2
(a) (b)
Figure 9.6 (a) G-type antiferromagnetic plane of BFO is perpendicular to the ferroelectric
polarized [111]c direction, where antiferromagnetic exist in both intra-plane and inter-plane.
(b) Magnetic structure including the Dzyaloshinskii–Moriya interaction. The two Fe3+ magnetic
moment (MFe1 and MFe2 ) are canted in the (111)c plane so that there is a resulting macroscopic
remnant magnetization M. Source: Wu et al. (2016). Reprinted with permission of Elsevier.
Bi
Fe
[001] 71
°
0
(a) (b)
0°
18
109
°
(c) (d)
Figure 9.7 (a) original polarization, (b) after 71∘ rotation, (c) after 109∘ rotation, and (c) after
180∘ rotation. Source: Adapted from Zhao et al. (2006).
9.5 ME Composite Materials 205
polarization can be switched by 71∘ to the direction shown in (b) and 109∘ switch-
ing to (c) and 180∘ switching to (d).
(a) (b)
(b) Unimorph
PVDF
Metglas
M
P
M
(c) T–L
9.6 Modeling the Interfacial Coupling in Multilayered ME Thin Film 207
ΔM = 16 emu/cm3
200
Magnetization (emu/cm3)
BaTiO3 matrix
100
50
300 320 340 360 380 400 420 440
Temperature (K)
composite films (Nan et al. 2005), which further proves that the substrate
clamping effect will destroy the mechanical coupling between piezoelectric and
magnetostrictive phase in the 2–2 heterostructured thin film.
500
400
300
ΔP3(μC/m2)
100
0
0 100 200 300 400
(a) H3(kA/m)
Composite
film with 1–3
nanostructure
Substrate
Composite film
with 2–2
structure
Substrate
(b)
Figure 9.11 (a) Theoretical ME coupling effect in nano and bulk systems with 1–3 and 2–2
structures, and (b) the corresponding schematic diagrams of 1–3 and 2–2 nanostructures.
Source: Adapted from Nan et al. (2005).
STO022
PZT022
CFO022
500 nm
(a) (b)
Figure 9.12 (a) Low-magnification TEM images of 11-layered PZT/CFO nanocomposite films,
(b) the selected area electron diffraction pattern including PZT, CFO, and STO substrate.
Source: Adapted from Zhang et al. (2010).
where Em and Ep are the electric field applied on the multilayers and the actual
field applied on each PZT layer, respectively; 𝜙P , 𝜙C , 𝜖 P , and 𝜖 C are the volume
fractions and dielectric constants of the PZT and CFO layers in the films. There-
fore, we have to increase the external field in the multilayers in order to maintain
the switching electric field in each PZT layer. It is easy to understand the signif-
icant enhancement of the coercive field in the films compared with that in the
single phase PZT film.
To modeling the number of layer-dependent polarization in the multilayered
thin films, the average residual stress suffered by the PZT phase in the epitaxial
multilayers can be estimated based on the XRD result. The in-plane a-axis lattice
parameters of PZT increase from 4.027 to 4.032, 4.035, and 4.041 Å, which
correspond to single PZT, 3-layered, 5-layered, and 11-layered films. The Young’s
modulus of the epitaxial PZT film is 92.6 GPa as reported elsewhere (Kanno
et al. 2004). Therefore the residual in-plane compressive stress suffered by the
corresponding epitaxial PZT films should be −520, −410, −330, and −204 MPa;
while for the cubic PZT bulk materials without residual stress, the lattice con-
stant a = 4.05 Å (Xu 1991). We adopt Landau–Devonshire’s phenomenological
thermodynamic formalism method to calculate the spontaneous polarization
210 9 Multiferroics: Single Phase and Composites
ΔG = 𝛼1 (P12 + P22 + P32 ) + 𝛼11 (P14 + P24 + P34 ) + 𝛼12 (P12 P22 + P22 P32 + P32 P12 )
+ 𝛼111 (P16 + P26 + P36 ) + 𝛼112 [P14 (P22 + P32 ) + P24 (P32 + P12 ) + P34 (P12 + P22 )]
1
+ 𝛼123 P12 P22 P32 − S11 (𝜎12 + 𝜎22 + 𝜎32 ) − S12 (𝜎1 𝜎2 + 𝜎2 𝜎3 + 𝜎3 𝜎1 )
2
1
− S44 (𝜎4 + 𝜎5 + 𝜎62 ) − Q11 (𝜎1 P12 + 𝜎2 P22 + 𝜎3 P32 )
2 2
2
− Q12 [𝜎1 (P22 + P32 ) + 𝜎2 (P32 + P12 ) + 𝜎3 (P12 + P22 )]
− Q44 (𝜎4 P2 P3 + 𝜎5 P3 P1 + 𝜎6 P1 P2 ) (9.6)
𝜎1 = 𝜎2 = −𝜎
𝜎3 = 𝜎4 = 𝜎5 = 𝜎6 = 0
P1 = P2 = 0 (9.7)
Therefore,
G = (𝛼1 + 2Q12 𝜎)P32 + 𝛼11 P34 + 𝛼111 P36 − (S11 + S12 )𝜎 2 (9.8)
dG
From the previously mentioned derivation and the condition dP3
= 0, we can
obtain
√
2 2
−𝛼11 ± 𝛼11 2 − 3𝛼111 (𝛼1 + 2Q12 𝜎)
PP = P3 = (9.9)
3𝛼111
Since the “-” in the equation is meaningless, it should be deleted.
In addition, according to the discussion of the “dilution” of the effect of ferro-
electric materials in a dielectric matrix in the BTO–STO–CTO superlattice (Lee
et al. 2005), when we consider the influence of the non-ferroelectric layer, the
electric displacement should be expressed by
( )
𝜖P 𝜖C 𝜙C 𝜖P
D= E+ 1− PP
𝜙 P 𝜖 C + 𝜙C 𝜖 P 𝜙P 𝜎C + 𝜙C 𝜎P
where D and P are the electric displacement and polarization of the multilayers.
Then the average polarization in the electrostatic model can be derived
PP
Pm = 𝜙C 𝜖P
(9.10)
1+ 𝜙P
⋅ 𝜖C
9.6 Modeling the Interfacial Coupling in Multilayered ME Thin Film 211
Table 9.1 Parameters used in calculating the polarization of the PZT film and PZT/CFO
multilayered heterostructures.
References
Apostolova, I. and Wesselinowa, J.M. (2008). Magnetic control of ferroelectric
properties in multiferroic BiFeO3 nanoparticles. Solid State Communications 147
(3): 94–97. Available at: http://www.sciencedirect.com/science/article/pii/
S0038109808002500.
Ascher, E. et al. (1966). Some properties of ferromagnetoelectric nickel-iodine
boracite, Ni3 B7 O13 I. Journal of Applied Physics 37 (3): 1404–1405. Available at:
https://doi.org/10.1063/1.1708493.
Astrov, D.N. (1960). The magnetoelectric effect in antiferromagnetics. Soviet Physics
Journal of Experimental and Theoretical Physics 11 (3): 708–709.
Baek, S.H., Jang, H.W., Eom, C.B. et al. (2010). Ferroelastic switching for nanoscale
non-volatile magnetoelectric devices. Nature Materials 9: 309. Available at:
http://dx.doi.org/10.1038/nmat2703.
Béa, H., Bibes, M., Barthélémy, A. et al. (2008). Mechanisms of exchange bias with
multiferroic BiFeO3 epitaxial thin films. Physical Review Letters 100 (1): 017204.
Available at: https://link.aps.org/doi/10.1103/PhysRevLett.100.017204.
Chu, Y.-H., Martin, L.W., Holcomb, M.B., and Ramesh, R. (2007). Controlling
magnetism with multiferroics. Materials Today 10 (10): 16–23. Available at:
http://www.sciencedirect.com/science/article/pii/S1369702107702419.
Curie, P. (1894). Sur la symétrie dans les phénomènes physiques, symétrie d’un
champ électrique et d’un champ magnétique. Journal of Physics: Theories and
Application 3 (1): 393–415. Available at: https://doi.org/10.1051/jphystap:
018940030039300.
Devan, R.S. and Chougule, B.K. (2007). Effect of composition on coupled electric,
magnetic, and dielectric properties of two phase particulate magnetoelectric
composite. Journal of Applied Physics 101 (1): 014109. Available at: https://doi
.org/10.1063/1.2404773.
Dzyaloshinskii, I.E. (1960). On the magneto-electrical effect in antiferromagnets.
Soviet Physics Journal of Experimental and Theoretical Physics 10 (3): 628–629.
Ederer, C. and Spaldin, N.A. (2005). Weak ferromagnetism and magnetoelectric
coupling in bismuth ferrite. Physical Review B 71 (6): 060401. Available at: https://
link.aps.org/doi/10.1103/PhysRevB.71.060401.
Eerenstein, W., Mathur, N.D., and Scott, J.F. (2006). Multiferroic and
magnetoelectric materials. Nature 442 (7104): 759–765. Available at: http://www
.ncbi.nlm.nih.gov/pubmed/16915279 (accessed 30 May 2014).
Heyer, O., Hollmann, N., Khomskii, D. et al. (2006). A new multiferroic material:
MnWO4 . Journal of Physics: Condensed Matter 18 (39): L471. Available at: http://
stacks.iop.org/0953-8984/18/i=39/a=L01.
Hoon Oh, S. and Jang, H.M. (1998). Enhanced thermodynamic stability of
tetragonal-phase field in epitaxial Pb(Zr,Ti)O3 thin films under a two-dimensional
compressive stress. Applied Physics Letters 72 (12): 1457–1459. Available at:
https://doi.org/10.1063/1.120609.
Hur, N., Park, S., Sharma, P.A. et al. (2004). Electric polarization reversal and
memory in a multiferroic material induced by magnetic fields. Nature 429: 392.
Available at: http://dx.doi.org/10.1038/nature02572.
References 213
Nan, C.-W., Liu, G., Lin, Y., and Chen, H. (2005). Magnetic-field-induced electric
polarization in multiferroic nanostructures. Physical Review Letters 94 (19):
197203. Available at: https://link.aps.org/doi/10.1103/PhysRevLett.94.197203.
Roginskaya, Y.E., Tomashpol’Skiı̌, Y.Y., Venevtsev, Y.N. et al. (1966). The nature of
the dielectric and magnetic properties of BiFeO3 . Soviet Physics Journal of
Experimental and Theoretical Physics 23: 47.
Schmid, H. (1994). Multi-ferroic magnetoelectrics. Ferroelectrics 162 (1): 317–338.
Available at: https://doi.org/10.1080/00150199408245120.
Seidel, J., Martin, L.W., He, Q. et al. (2009). Conduction at domain walls in oxide
multiferroics. Nature Materials 8: 229. Available at: http://dx.doi.org/10.1038/
nmat2373.
Siratori, K., Kohn, K., and Kita, E. (1992). Magnetoelectric effect in magnetic
materials. Acta Physica Polonica A 81 (4–5): 431–466.
Smolenskii, G.A. and Chupis, I.E. (1982). Ferroelectromagnets. Soviet Physics
Uspekhi 25 (7): 475–493. Available at: https://ufn.ru/en/articles/1982/7/b/.
Srinivasan, G., Rasmussen, E.T., Gallegos, J. et al. (2001). Magnetoelectric bilayer
and multilayer structures of magnetostrictive and piezoelectric oxides. Physical
Review B 64 (21): 214408. Available at: https://link.aps.org/doi/10.1103/PhysRevB
.64.214408.
Srinivasan, G., Rasmussen, E.T., and Hayes, R. (2003). Magnetoelectric effects in
ferrite-lead zirconate titanate layered composites: the influence of zinc
substitution in ferrites. Physical Review B 67 (1): 014418. Available at: https://link
.aps.org/doi/10.1103/PhysRevB.67.014418.
Van den Boomgaard, J., Terrell, D.R., Born, R.A.J., and Giller, H.F.J.I. (1974). An in
situ grown eutectic magnetoelectric composite material: Part I Composition and
unidirectional solidification. Journal of Materials Science 9 (10): 1705–1709.
Available at: https://doi.org/10.1007/BF00540770.
vanSuchtelen, J. (1972). Product properties: a new application of composite
materials. Philips Research Reports 27 (1): 28–37.
Wang, J., Neaton, J.B., Ramesh, R. et al. (2003). Epitaxial BiFeO3 multiferroic thin
film heterostructures. Science 299 (5613): 1719–1722.
Wong, C.K. and Shin, F.G. (2007). A simplified treatment of the Landau theory of
phase transitions for thin ferroelectric films. American Journal of Physics 76 (1):
31–38. Available at: https://doi.org/10.1119/1.2800353.
Wu, J., Fan, Z., Wang, J. et al. (2016). Multiferroic bismuth ferrite-based materials
for multifunctional applications: ceramic bulks, thin films and nanostructures.
Progress in Materials Science 84: 335–402. Available at: http://www.sciencedirect
.com/science/article/pii/S0079642516300585.
Xu, Y. (1991). Perovskite-type ferroelectrics: Part I. In: Ferroelectric Materials and
their Applications (ed. Y. Xu) In Y. B. T.-F. M. and their A. Xu, ed., 101–162.
North-Holland, Amsterdam: Elsevier. Available at: http://www.sciencedirect
.com/science/article/pii/B9780444883544500085.
Zavaliche, F., Yang, S.Y., Ramesh, R. et al. (2006). Multiferroic BiFeO3 films: domain
structure and polarization dynamics. Phase Transitions 79 (12): 991–1017.
Available at: https://doi.org/10.1080/01411590601067144.
References 215
Zavaliche, F., Zhao, T., Zheng, H. et al. (2007). Electrically assisted magnetic
recording in multiferroic nanostructures. Nano Letters 7 (6): 1586–1590.
Available at: https://doi.org/10.1021/nl070465o.
Zhai, J., Dong, S., Xing, Z. et al. (2006). Giant magnetoelectric effect in
metglas/polyvinylidene-fluoride laminates. Applied Physics Letters 89 (8): 083507.
Available at: https://doi.org/10.1063/1.2337996.
Zhang, J.X., Dai, J.Y., and Chan, H.L.W. (2010). Interfacial engineering and coupling
of electric and magnetic properties in Pb(Zr0.53Ti0.47)O3 /CoFe2 O4 multiferroic
epitaxial multilayers. Journal of Applied Physics 107 (10): 104105. Available at:
https://doi.org/10.1063/1.3386510.
Zhao, T., Scholl, A., Ramesh, R. et al. (2006). Electrical control of antiferromagnetic
domains in multiferroic BiFeO3 films at room temperature. Nature Materials 5:
823–829. Available at: http://dx.doi.org/10.1038/nmat1731.
Zheng, H., Wang, J., Lofland, S.E. et al. (2004). Multiferroic BaTiO3 -CoFe2 O4
nanostructures. Science 303 (5658): 661 LP–663. Available at: http://science
.sciencemag.org/content/303/5658/661.abstract.
Zhong, W.L., Wang, Y.G., Zhang, P.L., and Qu, B.D. (1994). Phenomenological study
of the size effect on phase transitions in ferroelectric particles. Physical Review B
50 (2): 698–703. Available at: https://link.aps.org/doi/10.1103/PhysRevB.50.698.
217
10
M M
P P
M M
L–T T–T
M
P M P
M
M P
Multi-L–T C–C
Zm Lm Cm φp : 1
Magnetoelastic Elastic–electric
coupling φm coupling φp
M
Piezoelectric plate
P
M
Figure 10.3 Schematic diagram of the ME laminate composite device with Terfenol-D
continuous fiber composite and piezoelectric plate. The arrows represent the magnetization
and polarization directions, respectively.
Measuring Hac
ME laminate
DC-coupled AC amplifier composite
Signal-in Signal-out
5
10 80
40 Hbias = 20 Oe
Impedance Hbias = 100 Oe
Phase angle 40 Hbias = 300 Oe
(ME)v (V/cmOe)
30 Hbias = 500 Oe
4
10
Hbias = 700 Oe
Z (Ω)
θ (°)
20
10
3
−40
10
−80
0
2
10
20.0k 40.0k 60.0k 80.0k 100.0k 120.0k 140.0k 0 20 40 60 80 100
(a) Frequency (Hz) (b) Frequency (Hz)
Figure 10.5 (a) Electrical impedance (Z) and phase angle (𝜃) spectra for a ME laminated
composite. (b) Magnetoelectric voltage coefficient (ME v ) as a function of frequency f for a ME
laminated composite tested under different dc bias magnetic field. Source: Adapted from
Cheung (2009).
2000 Hbias = 0 Oe
Hbias = 200 Oe
Hbias = 500 Oe
1600
Hbias = 800 Oe
Hbias = 1200 Oe
1200
V3 (mV)
Hbias = 1600 Oe
Hbias = 2000 Oe
800
400
0
0 2 4 6 8 10
H3 (Oe)
Figure 10.6 V ME (V 3 is the voltage across the piezoelectric plate) as a function of Hac with
f = 1 kHz for the ME laminated composite tested under different dc bias magnetic fields.
Source: Adapted from Cheung (2009).
also be identified in Figure 10.5b, but showing right shift under increased bias
field.
Figure 10.6 shows the V ME as a function of H ac with f = 1 kHz for the ME
laminated composite tested under different dc bias magnetic fields. It is appar-
ent that V ME shows excellent linear relationship with the measured range of H ac
under H bias . The linear response to H ac makes the ME laminated composite very
suitable for the application in ac magnetic field measurement.
During the operation of ME device, the magnetic bias field plays a very
important role that affects the sensitivity of the device. Figure 10.7 shows the
optimum H bias for the laminated composite device, where one can see that the
MEv increases with increasing H bias , but it finally reaches a maximum value
at H bias = 400 Oe and then starts to decrease from the maximum. This can be
understood by considering the fact that without bias field, the ac magnetic field
10.1 ME Composite Devices 221
3.5
@ 1 kHz
3.0
ME coefficient (αE) (V/cmOe)
2.5
2.0
1.5
1.0
0.5
0.0
0 200 400 600 800 1000 1200
Hbias (Oe)
22 mm
Stainless frame
3.0 MPa
0.15 0.15
0.10 0.10
0.05 0.05
0.00 0.00
0 20 40 60 80 100
Frequency (kHz)
Rap
V+
P FE – AFM
Electrode
Rp
V–
P FE – AFM
Electrode
Voltage
60 nm
1
δ V (μV)
Output signal 0
–1
3
δ H (Oe)
Read head
CFO
STO
BTO
Substrate
FM layer
FE layer
ME sensor 0
–3
δH dM
S dt
N S
N N S
S N
S N
N
S Iw
NS
S N
SN
N bK 0 0 1 1 0 1 1 1 1 1 0 1 1 0 1 0
Recording medium S aK 0 0 1 1 1 0 1 0 0 0 1 0 1 1 0 0
Figure 10.11 Schematic, TEM image (inset) and voltage output of an ME read head sensor
based on a bilayer CFO/BTO heterostructure grown on a single-crystal SrTiO3 (STO) substrate.
Source: Zhang et al. (2008). Reprinted with permission of AIP Publishing.
act not only as the magnetoelectric active layer but also a ferroelectric tunneling
barrier.
Another method, which has already been implemented in recording read
head application, is a junction working with magnetoelectric (ME) coupling. By
using strain as a media, in-plane magnetic field-induced strain in the magnetic
component due to magnetostrictive effect can be mechanically transferred
to the ferroelectric component and finally can induce an electric polarization
change, and thus voltage change, through the piezoelectric effect. Figure 10.11
demonstrates the ME coupling between the magnetostrictive CoFe2 O4 (CFO)
and ferroelectric BaTiO3 (BTO) (Zhang et al. 2008).
FM NM FM
I FE FE
FM NM FM
MTJ FTJ MFTJ
R R R
E
H
H-field E-field
Hc Ec Ec
Hc
Figure 10.12 Schematic illustration of artificial multiferroic tunnel junctions and its working
principle.
M
P
Ferromagnetic Ferroelectric
M barrier BTO
oxide LSMO
Substrate
STO
Aiming at constructing a MFTJ that can fulfill the constraint of less power
consumption, low fields in both ferroelectric and ferromagnetic switching is pre-
ferred. As an example of our work, Figure 10.13 shows a ferroelectric BTO tun-
nel barrier sandwiched by two electrodes composed of different ferromagnetic
materials La0.7 Sr0.3 MnO3 (LSMO) and Permalloy Ni0.81 Fe0.19 (NiFe). NiFe is a
well-known alloy that allows strong spin electron injection and has relatively
much lower coercive field, while LSMO has been predicted to have close to 100%
spin polarization. This structure is expected to form a low-switching-field spin
valve memory.
Figure 10.14 shows resistance–voltage (R–V ) characteristics of non-volatile
resistance switching from this MFTJ device. Unlike RRAM-resistive ran-
dom access memory (ReRAM), no electroforming at higher voltage is
needed before the observation of hysteretic resistance changes. During the
measurements, pulse train writing voltages are applied in the sequence of
0 V → 1.5 V → 0 V → −1.5 V → 0 V with a step of 0.2 V, while reading voltage is
remained constant as 0.1 V. With the top electrode grounded, a positive pulse of
226 10 Device Application of Multiferroics
80
70
60
50
Resistance (kΩ)
40
30
20
10
voltage (+V ) is applied on the bottom LSMO electrode, which drives the BaTiO3
ferroelectric polarization pointing to the NiFe electrode and sets the junction to
ON state (low resistance state). In contrast to the ON state, a negative voltage
pulse (−V ) drives the polarization to the opposite direction and sets the junction
to OFF state (high resistance state).
Magnetic characteristics of the NiFe and LSMO electrodes are shown in
Figure 10.15, where two distinct M–H hysteresis loops present in both layers.
NiFe has almost the same magnetization as LSMO; however, the coercive field
of LSMO is 20 Oe, while it is only 5.0 Oe for NiFe in the film plane. The different
coercive fields for these two dissimilar electrodes determine that LSMO acts as
a hard magnet and NiFe electrode acts as a soft magnet in the spin valve. This
result suggests that two resistance states can be achieved by interchanging the
magnetization orientations using external magnetic field.
Figure 10.16 demonstrates the coexistence of TMR and TER in an artificial
MFTJ. By placing the sample in the middle of two electromagnets, alternative
magnetic field is applied to the junction parallel to the film plane. Then, the junc-
tion (ferroelectric barrier) was poled by 1.5 V and −1.5 V for poling upward and
downward, respectively. Among the four stages, two resistance states are formed
due to the electric polarization switching and the other two are due to the parallel
and antiparallel magnetization configurations.
By sweeping the external magnetic field from −100 Oe to 100 Oe and returning
to −100 Oe, two distinct resistance states can be clearly observed for the two
ferroelectric polarization states of up and down, i.e. there are four distinct
10.3 Memory Devices Based on Multiferroic Tunneling 227
800
LSMO
600 NiFe
Magnetization (μemu/cm2)
400
200
–200
–400
–600
–800
Figure 10.15 Magnetism of NiFe and LSMO showing the hysteresis loops. Source: Adapted
from Yau et al. (2015).
175.4
D B
175.2
175.0 A C
Resistance (kΩ)
174.8
174.6
174.4
10.54
A′ C′
10.53 D′ B′
10.52
–100 –50 0 50 100
(a) Magnetic field (Oe)
A B C D
A′ B′ C′ D′
(b)
Figure 10.16 (a) Resistance changes with different magnetic fields at 8 K showing four-state
memory tested on the NiFe/BaTiO3 /LSMO structure. (b) The corresponding schematic diagram
for magnetization and ferroelectric polarization orientations. Source: Adapted from Yau et al.
(2015).
References
Bibes, M. and Barthélémy, A. (2008). Multiferroics: towards a magnetoelectric
memory. Nature Materials 7 (6): 425–426. Available at: http://www.ncbi.nlm.nih
.gov/pubmed/18497843%5Cnhttp://www.nature.com/doifinder/10.1038/
nmat2189.
Cheung, K. (2009). Ultrasonic Transducer Equipped with a magnetoelectric sensor
for weld quality monitoring. Hong Kong Polytechnic University. Available at:
https://theses.lib.polyu.edu.hk/handle/200/3843.
DeTeresa, J.M., Barthélémy, A., Fert, A. et al. (1999). Role of metal-oxide interface in
determining the spin polarization of magnetic tunnel junctions. Science 286
(5439): 507–509. Available at: http://www.sciencemag.org/cgi/doi/10.1126/
science.286.5439.507.
Dong, S., Li, J.-F., and Viehland, D. (2006). Magnetoelectric coupling, efficiency, and
voltage gain effect in piezoelectric-piezomagnetic laminate composites. Journal of
Materials Science 41 (1): 97–106. Available at: https://doi.org/10.1007/s10853-
005-5930-8.
Eerenstein, W., Mathur, N.D., and Scott, J.F. (2006). Multiferroic and
magnetoelectric materials. Nature 442 (7104): 759–765. Available at: http://www
.ncbi.nlm.nih.gov/pubmed/16915279.
Gajek, M., Bibes, M., Fusil, S. et al. (2007). Tunnel junctions with multiferroic
barriers. Nature Materials 6 (4): 296–302. Available at: http://www.ncbi.nlm.nih
.gov/pubmed/17351615.
Julliere, M. (1975). Tunneling between ferromagnetic films. Physics Letters A 54 (3):
225–226.
Lam, K.H., Lo, C.Y., Dai, J.Y. et al. (2011). Enhanced magnetoelectric effect in a
stress-biased lead magnesium niobate-lead titanate single crystal/Terfenol-D
alloy magnetoelectric sensor. Journal of Applied Physics 109 (2): 024505. Available
at: https://doi.org/10.1063/1.3536636.
Nan, C.-W., Bichurin, M.I., Dong, S. et al. (2008). Multiferroic magnetoelectric
composites: historical perspective, status, and future directions. Journal of
Applied Physics 103 (3): 031101. Available at: https://doi.org/10.1063/1.2836410.
Wang, Y., Or, S.W., Chan, H.L.W. et al. (2008). Enhanced magnetoelectric effect in
longitudinal-transverse mode Terfenol-D/Pb(Mg1/3 Nb2/3 )O3 –PbTiO3 laminate
230 10 Device Application of Multiferroics
composites with optimal crystal cut. Journal of Applied Physics 103 (12): 124511.
Available at: https://doi.org/10.1063/1.2943267.
Yau, H.M., Yan, Z.B., Dai, J.Y. et al. (2015). Low-field switching four-state nonvolatile
memory based on multiferroic tunnel junctions. Scientific Reports 5: 12826.
Available at: http://www.nature.com/doifinder/10.1038/srep12826.
Zhang, Y., Li, Z., Nan, C.W. et al. (2008). Demonstration of magnetoelectric read
head of multiferroic heterostructures. Applied Physics Letters 92 (15): 152510.
Available at: https://doi.org/10.1063/1.2912032.
231
11
Shape memory alloys (SMAs), or generally being called shape memory materials
(SMMs) since some polymers also have shape memory characteristics, can be
defined as a class of materials having the ability to “memorize” or retain their
previous shape in response to stimulus such as thermal and mechanical or even
magnetic variations. To be a member in the family of “ferroic,” SMA is charac-
terized as ferroelastic exhibiting hysteresis loop in the curves of strain–stress,
strain–temperature, or strain–magnetic field. Analogous to the other two family
members of ferromagnetic and ferroelectric, ferroelastic materials also present
domain structures, but instead of electrical and magnetic orderings, the domains
are segregated regions of martensite variants. Not only SMA and magnetic SMAs
but also many ferroelectric and ferromagnetic materials are ferroelastic; however,
the desired devices as introduced in previous chapters are mainly related to ferro-
electric and ferromagnetic switching. Many metal oxide materials such as SrTiO3
and LaAlO3 also exhibit ferroelasticity, but the range is relative too small for
engineering application. One of the first full ferroelastic hysteresis loops (Figure
11.1) was seen in 1976 for the prototypic material Pb3 (PO4 )2 (Salje 1993; Salje and
Hoppmann 1976) and was introduced in the review of ferroelastics (Salje 2012).
Pb b
P Strain (10–4)
c
Stress (bars)
–4 –3 –2 –1 1 2 3 4
Figure 11.1 Ferroelastic hysteresis and atomic switching in Pb3 (PO4 )2 . The structural
mechanism for ferroelasticity originates from the shift of the Pb atom away from the center of
its oxygen cage toward one pair of oxygen ions. This leads to a monoclinic distortion of the
crystal structure that can be inverted or rotated (in the direction of the arrows) under external
stress. Source: Reprinted with permission from Salje (2012).
Ni
Ti
(a) (b)
Figure 11.2 (a) Austenite phase (B2) and (b) martensite lattice structure (B19’) of
nickel–titanium, where the plane outlined by dash line represents the {100} planes in the
primitive cell of the martensite structure.
Figure 11.3 Lattice cell of NiTi and four martensite variants developed from one face-diagonal
plane.
Cooling
Heat flow
Mf Ms
As Af
Heating
Temperature
Figure 11.4 A schematic diagram of a typical differential scanning calorimeter (DSC) curve for
NiTi shape memory alloy showing presence of latent heat during phase transformation.
Mf As
100
Martensite (%)
Length
ΔL
0
Ms Af
Temperature
Figure 11.5 Typical transformation–temperature curve of NiTi SMA under constant stress as it
is cooled and heated showing hysteresis.
Stress
Stress-induced strain
Mf
Strain
Cooling
Af
Heating
Temperature
He
ati
ng
Stress/load
Loading
Cooling
Heating
(a) Temperature
He
ati
ng
Stress/load
Loading
Cooling
Heating
(b) Temperature
Figure 11.7 Illustration of (a) one-way and (b) two-way memory effects.
Loading
Unloading
Strain
SMA has also been used in fire security and protection systems, where the SMA
should have an immediate action to shut down or open the valve in the presence
of increased heat from fire. This can greatly decrease leakage-induced problems
in industries that involve petrochemicals, semiconductors, pharmaceuticals, and
large oil and gas boilers.
Many new applications of SMA are being developed, and even polymer-based
SMMs are emerging. These are beyond the scope of this book.
Ni Mn Ga
[110]
[110]
[001]
[001]
– –
[11 0] [11 0]
– –
[11 0] [11 0]
Figure 11.10 Illustrations of Ni2 MnGa structured models: (a) 3D cubic structure in the
high-temperature austenitic phase with two kinds of stacking (110) plane in cubic phase.
Five-layered martensite resulted from shear on these {110}c planes in the (b) cubic phase and
(c) modulation. Source: Adapted from Ge (2007).
a
a
a c
Variant I
c
Variant II
(a) (b)
Figure 11.11 Schematic illustrations of (a) martensitic transformation and (b) twinned
structure in two dimensions. Arrows indicate the orientations of the variants (or easy axis).
Source: Adapted from Enkovaara et al. (2004).
[100] [001]
20 μm
[001] [100]
Micro-scale
Large strain
Single variant I
H=0
(a)
Applying magnetic field
Twin variants
Hc < H < Hs
Single variant II
H ≥ Hs
(c)
Figure 11.13 Schematic illustrations of a Ni–Mn–Ga single crystal under increasing applied
magnetic field levels. Schematics (a)–(c) depict the respective microstructural arrangements of
martensitic variants and magnetic domains. Source: Adapted from Kiefer and Lagoudas (2005).
5
Strain (%)
3 c
c
a
2 a
0
0 100 200 300 400 500 600 700
H (kA/m)
Figure 11.14 MFIS dependence of applied magnetic field illustrates the change of twin
variants. Source: Adapted from Sozinov et al. (2003).
Figure 11.14 shows the relationship between MFIS and the corresponding
change of twin variants; where the coercive field H c and saturation field H s are
about 250 and 580 kA/m, respectively, and the maximum MFIS is about 5.5%. In
addition, the maximum strain is predictable from the theoretical lattice constant
with the equation
𝜀0 = (1 − c∕a)
where c and a are the short (or easy axis) and long (or hard axis) lattice constants,
respectively.
Third, data presented in MFIS curves indicate that a relatively large critical field
is needed to initiate shape transformation. Consequently, this relatively large crit-
ical field will increase the size of the drive coil/core, which significantly limits the
application in realizing miniature active systems.
Fourth, the output mechanical stress by Ni–Mn–Ga single crystals driven by
MFIS is actually very small (6–10 MPa). This means that the actuation stress in
practical application is small, which is difficult in meeting the requirement for
large actuation force.
Vac
Vind PMN–PT
0.9
(b)
MP MAPT AP
0.6 Ms
αB (G/V)
Af
0.3
Heating
Mf As Cooling
0.0
20 25 30 35 40 45 50 55 60
T (°C)
Figure 11.15 (a) Schematic diagram of the proposed heterostructure. (b) CME coefficient
(𝛼 B ) dependence on temperature at an applied ac electric voltage V ac of 20 V peak, a
frequency f of 1 kHz, and a magnetic bias field Hbias of 150 Oe upon heating and cooling.
Source: Zeng et al. (2010). Reprinted with permission of AIP Publishing.
244 11 Ferroelasticity and Shape Memory Alloy
References
Burchett, S. (2019). Global Nitinol Shape Memory Alloy Market 2019 Top Company
Profiles, Competition Status, Trends, Industry Growth, And Forecast 2019–2025.
themarketfact.com. Available at: https://themarketfact.com/2019/05/02/global-
nitinol-shape-memory-alloy-market-2019/ (accessed 21 August 2019).
Enkovaara, J., Ayuela, A., Zayak, A.T. et al. (2004). Magnetically driven shape
memory alloys. Materials Science and Engineering: A 378 (1): 52–60. Available at:
http://www.sciencedirect.com/science/article/pii/S0921509303015181.
Ge, Y. (2007). The Crystal and Magnetic Microstructure of Ni–Mn–Ga Alloys.
Helsinki University of Technology.
Ge, Y., Heczko, O., Söderberg, O., and Lindroos, V.K. (2004). Various magnetic
domain structures in a Ni–Mn–Ga martensite exhibiting magnetic shape
memory effect. Journal of Applied Physics 96 (4): 2159–2163. Available at: https://
doi.org/10.1063/1.1773381.
Henry, C.P., Bono, D., Feuchtwanger, J. et al. (2002). AC field-induced actuation of
single crystal Ni–Mn–Ga. Journal of Applied Physics 91 (10): 7810–7811.
Available at: https://aip.scitation.org/doi/abs/10.1063/1.1449441.
James, R.D., Tickle, R., and Wuttig, M. (1998). Large field-induced strains in
ferromagnetic shape memory materials. Materials Science and Engineering. A,
Structural Materials: Properties, Microstructure and Processing 273–275:
320–325.
Kiefer, B. and Lagoudas, D.C. (2005). Magnetic field-induced martensitic variant
reorientation in magnetic shape memory alloys. Philosophical Magazine 85
(33–35): 4289–4329. Available at: https://doi.org/10.1080/14786430500363858.
Salje, E.K. (1993). Phase Transitions in Ferroelastic and Co-elastic Crystals.
Cambridge, UK: Cambridge University Press.
Salje, E.K.H. (2012). Ferroelastic materials. Annual Review of Materials Research 42
(1): 265–283. Available at: https://doi.org/10.1146/annurev-matsci-070511-
155022.
Salje, E.K. and Hoppmann, G. (1976). Direct observation of ferroelasticity in
Pb3 (PO4 )2 -Pb3 (VO4 )2 . Materials Research Bulletin 11 (12): 1545–1549. Available
at: http://www.sciencedirect.com/science/article/pii/0025540876901070.
Sozinov, A., Likhachev, A.A., Lanska, N. et al. (2002). Giant magnetic-field-induced
strain in NiMnGa seven-layered martensitic phase. Applied Physics Letters 80
(10): 1746–1748. Available at: https://doi.org/10.1063/1.1458075.
Sozinov, A., Likhachev, A.A., Lanska, N. et al. (2003). Effect of crystal structure on
magnetic-field-induced strain in Ni–Mn–Ga. Proc. SPIE 5053, Smart Structures
and Materials 2003: Active Materials: Behavior and Mechanics 5053: 586–594.
References 245
Suorsa, I., Pagounis, E., and Ullakko, K. (2004). Magnetic shape memory actuator
performance. Journal of Magnetism and Magnetic Materials 272–276 (3):
2029–2030. Available at: http://www.sciencedirect.com/science/article/pii/
S0304885303024922.
Tickle, R. and James, R.D. (1999). Magnetic and magnetomechanical properties of
Ni2MnGa. Journal of Magnetism and Magnetic Materials 195 (3): 627–638.
Available at: http://www.sciencedirect.com/science/article/pii/
S0304885399002929.
Tickle, R., James, R.D., Shield, T. et al. (1999). Ferromagnetic shape memory in the
NiMnGa system. IEEE Transactions on Magnetics 35 (5): 4301–4310.
Ullakko, K., Huang, J.K., Kantner, C. et al. (1996). Large magnetic-field-induced
strains in Ni2MnGa single crystals. Applied Physics Letters 69 (13): 1966–1968.
Zeng, M., Or, S.W., and Chan, H.L.W. (2010). Effect of phase transformation on the
converse magnetoelectric properties of a heterostructure of Ni49.2 Mn29.6 Ga21.2
and 0.7PbMg1/3 Nb2/3 O3-0.3 PbTiO3 crystals. Applied Physics Letters 96 (18):
182503. Available at: https://doi.org/10.1063/1.3427388.
247
Index
toroidal domains 99 x
up and down polarization domains X-ray diffraction 38–40, 96, 110
101
y
w Young’s modulus 127, 128, 209
wake-up effect 110, 111