University of Southern California Dissertations and Theses

Development of high-frequency (~100mhz) PZT thick-film ultrasound transducers and arrays

(USC Thesis Other)

Development of high-frequency (~100mhz) PZT thick-film ultrasound transducers and arrays

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Copyright 2009 Dawei Wu

DEVELOPMENT OF HIGH-FREQUENCY (~100MHZ)
PZT THICK-FILM ULTRASOUND TRANSDUCERS AND ARRAYS

by

Dawei Wu

___________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOMEDICAL ENGINEERING)

December 2009

ii
ACKNOWLEDGMENTS
I would like to express my sincere gratitude to my advisor Professor K. Kirk Shung
and Professor Qifa Zhou, for their guidance, suggestions and encouragement throughout
the course of this research. I am truly indebted to them with a research assistantship
throughout the duration of my studies at the Resource Center for Medical Ultrasonic
Transducer Technology at the University of Southern California. It has been a great
pleasure to be there. I have learned a great deal on ultrasound imaging, transducer design
and fabrication in this excellent laboratory.
I am grateful to my thesis committees: K. Kirk Shung, Qifa Zhou, Jesse Yen,
Satwindar Sadhal and Yong Chen for advising me on the preparation of this dissertation
and reviewing my thesis manuscript.
My work would not have been realized without the help of Changgeng Liu and Frank
Djuth from Geospace Research, Inc and visiting student Benpeng Zhu from Wuhan
University, China. They provided valuable help on PZT film preparation and dry-etching
fabrication process. They also contributed to my knowledge of ultrasonic array design
and fabrication through helpful discussions.
I would also like to thank the staff, graduate students at the resource center, especially
our manager Jon Cannata, senior technician Jay Williams, research faculty Changhong
Hu, and our budget analyst Peter Lee. They supported me throughout my study and made
my stay at the center an enjoyable experience.
Above all, a heartfelt expression of gratitude to my parents for all their support and
encouragement.

iii
TABLE OF CONTENTS

ACKNOWLEDMENTS ii

LIST OF TABLES v

LIST OF FIGURES vi

ABSTRACT ix

Chapter 1. INTRODUCTION TO HIGH-FREQUENCY ULTRASONIC
IMAGING
1
1.1 High-frequency ultrasonic imaging 1
1.2 High-frequency single element transducers and arrays 4
1.2.1 KLM model 5
1.2.2 Mechanical matching 7
1.2.3 Electrical matching 10
1.2.4 Single element transducers 11
1.2.5 Linear arrays 12
1.2.6 Issues in high frequency ultrasonic imaging 14
1.3 Scopes of the research 17

Chapter 2. INTRODUCTION TO PIEZOELECTRIC MATERIALS 19
2.1 Introduction 19
2.1.1 Phase transitions and perovskite structure 21
2.1.2 PZT solid solution and the morphotropic phase boundary 22
2.1.3 Poling mechanism of ferroelectrics 24
2.1.4 Ferroelectric hysteresis loop 25
2.2 Piezoelectric materials 27

Chapter 3. PZT THICK FILMS FABRICATION 35
3.1 Introduction to PZT thick-films 35
3.2 Fabrication of PZT thick-films 38
3.2.1 Sol-gel infiltration 40
3.2.2 Milling time control 42
3.2.3 Powder to solution mass ratio 44
3.2.4 Summary 50
3.3 Characterization of the PZT thick-film 51
3.4 Poling the PZT thick films 57

iv

Chapter 4. FABRICATION OF PZT THICK FILM SINGLE-ELEMENT
TRANSDUCERS AND KERFLESS ARRAYS
59
4.1 Fabrication of PZT single-element transducers 60
4.1.1 Modelling 60
4.1.2 Fabrication process 61
4.2 Fabrication of PZT kerfless arrays 62
4.2.1 Modelling 63
4.2.2 Fabrication process 64
4.3 Results and discussion 67
4.3.1 Single element transducer 71
4.3.2 Kerfless arrays 73
4.3.3 Ultrasonic imaging with film single-element transducer 78

Chapter 5. FABRICATION OF PZT THICK FILM KERFED ARRAYS 81
5.1 Dry-etching technique 81
5.2 Modelling 82
5.3 Fabrication process 83
5.4 Results and discussion 90

Chapter 6. SUMMARY AND FUTURE WORK 93
6.1 Summary of results 93
6.2 Suggestions for future work 95

BIBLIOGRAPHY 97

v

LIST OF TABLES

Table 1.1 Frequency dependence of α applies at the range 1-10MHz of some
tissues relevant to ultrasonic imaging
15

Table 2.1 Summary of the ZnO, AIN and PZT films piezoelectric and
dielectric properties
32

Table 2.2 Properties of different piezoelectric materials 34

Table 3.1 Important properties of PZT film and PZT-5H 56

Table 4.1 Modelling and measured results of representative PZT thick film
and PZT-5H ceramic elements.
76

vi
LIST OF FIGURES

Fig. 1.1 Plot of axial resolution versus frequency for transducers with
various bandwidth (a), and plot of lateral resolution versus
frequency for transducers with various f-number (b).
2

Fig. 1.2 KLM electrical equivalent model. 5

Fig. 1.3 Simplified model of an ultrasonic transducer. 7

Fig. 1.4 Serious (a) and parallel (b) equivalent circuit models for a single
element transducer at its resonance
11

Fig. 1.5 Attenuation of variety tissues in the frequency range from 10 to
100 MHz, from Foster (2000)
16

Fig. 2.1 Phase diagram of Pb(Zr, Ti)O
3
solid solution from Xu (1991) 22

Fig. 2.2 Ferroelectric with random orientation of grains before and after
poling, from Xu (1991)
24

Fig. 2.3 Ferroelectric (P-E) hysteresis loop, from Damjanovic (1998) 25

Fig. 3.1 Flowchart of the PZT film fabrication process. 40

Fig. 3.2 SEM pictures of PZT thick films without (a) and with (b) sol-gel
infiltration process.
41

Fig. 3.3 Particle size distribution of the PZT composite solution after
10-hour (a), 20-hour (b), and 30-hour (c) milling.
43

Fig. 3.4 Hysteresis loop of thick films with variable solution-to-powder
mass ratios.
45

Fig. 3.5 Dielectric properties of thick films with variable solution-to-
powder mass ratios.
45

Fig. 3.6 Dielectric constant and remanent polarization values as a function
of solution-to-powder mass ratio.
46

Fig. 3.7 e
31.f
values of the PZT thick films with variable solution-to-powder
mass ratios.
48

Fig. 3.8 Variation of film dielectric constants with film densities. 49

vii
Fig. 3.9 SEM images of top view (a) and cross-section view (b) of a PZT
film
51

Fig. 3.10 XRD pattern of a PZT thick film 52

Fig. 3.11 Hysteresis loop of a PZT thick film 53

Fig. 3.12 Dielectric properties of a PZT thick film 53

Fig. 3.13 SEM pictures of a PZT film surface after removing from the
silicon by KOH
54

Fig. 3.14 Measured electric impedance curves of a PZT thick film with (a)
and without (b) silicon substrate
55

Fig. 3.15 Electric field effects and temperature effects in poling process 58

Fig. 4.1 Modelling pulse-echo response of the PZT film single element
transducer
61

Fig. 4.2 Structure of the PZT film single element transducer 61

Fig. 4.3 Pattern of the PZT film linear array 62

Fig. 4.4 Modelling pulse-echo response of PZT film (a) and PZT-5H
ceramic kerfless array elements
63

Fig. 4.5 Fabrication process for PZT film kerfless arrays 66

Fig. 4.6 Picture of top electrodes of the PZT film linear array with
insulation layer underneath
66

Fig. 4.7 Pulse plots of 5900PR with 1µJ (a), 2 µJ (b) and 4 µJ (c) settings 67

Fig. 4.8 Pulse plots of 5910 with Low (a) and HIGH (b) energy settings 69

Fig. 4.9 Avtech AVB2-TC-C monocycle generator’s pulse plot 70

Fig. 4.10 Pulse-echo test set-up 71

Fig. 4.11 Measured pulse-echo plots of a PZT film sing-element transducer
before (a) and after (b) depositing parylene as a matching layer
72

Fig. 4.12 Measured pulse-echo response of PZT film (a) and PZT-5H
ceramic kerfless array elements
74

viii

Fig. 4.13 Pulse-echo (a), insertion loss (b) and cross-talk (c) plots of a
representative kerfless array element with parylene matching
76

Fig. 4.14 Ultrasonic image of wire targets with 12-dB (a) and 6-dB (b)
dynamic range
78

Fig. 4.15 Ultrasonic image of a normal porcine eyeball 80

Fig. 4.16 Ultrasonic image of a normal human skin 80

Fig. 5.1 FIELD II (a) and PZFLEX (b) modelling of the linear array 83

Fig. 5.2 SEM pictures of a linear array dry-etched from a PZT film 85

Fig. 5.3 Photography of the array before (a) and after (b) epoxy filling 86

Fig. 5.4 Layout (a) and photography (b) of the flexible circuit 87

Fig. 5.5 Layout (a) and photography (b) of the PCB connector board 87

Fig. 5.6 Top view (a) and bottom view (b) of the kerfed array before
interconnection
89

Fig. 5.7 Picture of the array after interconnection 89

Fig. 5.8 Measured pulse-echo of a kerfed array element 90

Fig. 5.9 Measured element uniformity of the kerfed array 91

Fig. 5.10 Measured cross-talk between adjacent elements 92

Fig. 5.11 Measured insertion-loss of a representative array element 92

ix

ABSTRACT
Fabrication of high-frequency (30 MHz -50 MHz) ultrasonic linear arrays is still a
challenge. The task is even more difficult to build arrays at a frequency higher than 100
MHz, which has the potential to provide more detail skin texture for early diagnosis of
melanoma or to image small objects such as stem cells. Integrating of PZT films into
MEMS micro-machined is a potential solution to such high-frequency applications. This
thesis presents the development of high-frequency (~100 MHz) PZT thick-film
transducers and arrays.
PZT thick films thicker than 10 µm were fabricated with a composite spin-coating
method. The functional properties of the films were improved by optimization of
composite solution ball-milling time, PZT sol-gel to powder mass ratio and sol-gel
infiltration process. Films suitable for high frequency ultrasonic transducer applications
were produced and characterized. The films were found to have a remanent polarization
of 37 µC/cm
2
, a dielectric constant of 1250 (at 1 kHz), and electromechanical coupling
coefficient (k
t
) of 0.34.
High-frequency (~100 MHz) single-element transducers and linear arrays were
afterwards fabricated with the PZT films. The single-element transducer was constructed
by transferring the film from silicon substrate to E-solder backing material. The finished
transducer was found to have a center frequency of 120 MHz and a bandwidth of 60%
with a layer of parylene. Ultrasonic images of the porcine eyeball and normal human skin
were also successfully acquired with the transducer. 32-element kerfless high-frequency
linear array was fabricated with a 12-µm PZT film by photolithography. The array had a

x
center frequency of 120 MHz, a bandwidth of 60% with parylene matching and an
insertion loss of 41 dB. Performance of the array was compared to a PZT-5H ceramic
sheet kerfless array fabricated with an identical array pattern.
The development of 32-elment kerfed arrays was also presented. DRIE dry-
etching technique was investigated to produce an array with sharp element edges (profile
angle > 85º) and narrow kerf ( ~12 µm). The kerfed array shows similar performance
with kerfless one, except the crosstalk is ~5 dB lower.

1

CHAPTER 1
INTRODUCTION TO HIGH-FREQUENCY ULTRASONIC
IMAGING

1.1 High Frequency Ultrasonic Imaging
An ultrasonic image is constructed from echoes returned from reflecting or
scattering targets. The positional information of the image depends on the measurement
of the time that elapses between the transmitted pulses and received echoes. The
brightness of the image depends on the amplitude of the echoes.
One of the most important characterizations of an ultrasonic image is its spatial
resolution which will determine how small a target can be differentiated from
surrounding objects. The spatial resolution of an ultrasound image depends on the
profiles of the ultrasonic beam and pulse produced from a transducer. Axial spatial
resolution (in depth) is determined by the duration of the ultrasound pulse within -6-dB of
the maxima. Mathematically it is given by
) 2 (
c
axial
f BW
c
R
× ×
= , (1.1)
where c is the sound speed in the medium detected, BW is the -6-dB percent bandwidth of
a transducer, and f
c
is the center frequency of a transducer.
Lateral spatial resolution (in elevation and azimuth direction) is conventionally
defined as the beam profile in the lateral direction within -6-dB of the maximum which
can be expressed as

2
λ × =
#
f R
lateral
, (1.2)
where
#
f is f-number of a transducer defined as the ratio of focal distance to diameter of
a transducer, λ is the wavelength in the medium.
Fig1.1 (a) shows the plot of axial resolution versus frequency for transducers with
percent bandwidth (-6 dB) of 25%, 50% and 75% in the human soft tissues. The curves of
lateral resolution versus frequency for transducers with f-number of 1, 2 and 6 in the
human soft tissues are plot in Fig1.1 (b). The plots suggest that transducer’s frequency is
the primary factor determining both the axial and the lateral resolutions of an ultrasonic
image: the higher the transducer’s frequency, the better the spatial resolution could be
obtained.
10
0
10
1
10
2
10
3
10
0
10
1
10
2
10
3
10
4
Frequency(MHz)
Axial resolution (um)

BW = 25%
BW = 50%
BW = 75%

(a)

3
10
0
10
1
10
2
10
3
10
0
10
1
10
2
10
3
10
4
Frequency(MHz)
Lateral resolution (um)

f # = 1
f # = 2
f # = 6

(b)

Fig 1.1 Plot of axial resolution versus frequency for transducers with percent
bandwidth of 25%, 50% and 75% (a), and plot of lateral resolution versus frequency for
transducers with f-number of 1, 2 and 6 (b).

In recent 20 years, ultrasound imaging of human skin, eyes and vascular anatomy
and small animals has been extensively studied. Conventional clinical ultrasound imaging
systems, which are operated at between 1 MHz and 10 MHz, cannot meet requirements
of these emerging applications which require finer spatial resolution of images. As just
discussed, to obtain improved spatial resolution of ultrasonic images, the ultrasound
imaging systems ultimately have to be operated at a higher frequency. In the middle of
80s, Foster et al (1989), developed an ultrasound backscatter microscopy (abbreviated as
UBM) which is a high-frequency ultrasound imaging system operating at frequency
beyond 30 MHz. Since then, high-frequency ultrasound imaging has been a very active
area of research. Foster et al (2000) demonstrated assessment of cardiac structures and
functions in mouse embryos. Hu et al. (2006) developed a high frequency linear array

4
ultrasonic imaging system. Sun et al. (2008) recently published the work of in vivo
cardiac imaging of adult zebra fish using high frequency ultrasound (45-75 MHz). High
frequency ultrasonic imaging brings challenges not only in design of imaging system, but
also in development of ultrasonic transducers. Design and fabrication of the ultrasonic
transducers suitable for high frequency medical imaging applications is the interest of this
research.

1.2 Introduction to Single-element Transducers and Linear Arrays
To produce a two-dimensional ultrasound image, ultrasound beam need to be
scanned across the imaging plane either by mechanical linear scanning of a single-
element transducer or by electronically scanning of an array. Mechanical scanning
involves a very simple system and transducer design. But wobble of the transducer during
scanning and image qualities are its main drawbacks. Moreover, mechanical scanning
normally needs a very bulky positioning equipment to control and scan the transducer.
Linear array scanning, on the other hand, allows the ultrasonic beam to be electronically
steered and focused. It can provide high-quality real-time images. But complexity and
cost of the linear array and its imaging system are always great concerns.
Both the single-element transducer and the linear array share the same criteria as a
good medical transducer: an ideal transducer for ultrasonic imaging would be perfectly
matched to loading medium (usually human tissues), have high efficiency as transmitter
and high sensitivity as a receiver, a wide dynamic range and a wide frequency response.
To achieve these criteria, both mechanical matching and electric matching should be

5
carefully considerated. In the following paragraphs of this section, one-dimensional KLM
model by Krimholtz et al (1970) is first introduced, which provide useful starting point
for optimizing transducers’ performance; mechanical matching which includes front
matching and rear backing and electric matching are briefly reviewed; design
considerations of the single-element transducer and the linear array are also discussed.

1.2.1 KLM Model
Popular 1-D transducer models include Mason model, the Redwood model and
KLM model. Among them, the KLM is more physically intuitive, thus is widely used.
This model divides a piezoelectric element into two halves, each represented by an
acoustic transmission line. The acoustic transmission line serves as a secondary circuit,
which is linked with an electric primary circuit by an ideal transformer as shown below in
Fig. 1.2.

Fig 1.2 KLM electrical equivalent model

Z
c
, c
p
, d/2
Z
c
, c
p
, d/2
Φ:1
C
o

C
’

Back acoustic port Front acoustic port
Electric port

6
The component and constant used in the model are list below
A = transducer area
d = thickness of the piezoelectric material
ρ = density of the piezoelectric material
c = the speed of sound in the piezoelectric material
c
Z
= radiation impedance of the piezoelectric layer (=ρcA)
ε
∗
= complex clamped dielectric permittivity
33
e
= piezoelectric constant
D
c
= elastic constant of piezoelectric layer
S
r
ε
= clamped dielectric constant
d A C /
*
0
ε =
, (1.3)
2
2
33
D S t
r
e
k
c ε
=
, (1.4)
d
c π
ω
2
0
=
, (1.5)
0 0
( )
t
c
k
C Z
π
ω
Φ=
, (1.6)
' 2 1
0 0
/ (sinc( / ))
t
C C k ω ω
−
=−
, (1.7)
With the KLM model, the electric impedance of the transducer is given as
)) ( (
1 1
2 1
2
2 1
'
0
Z Z
Z Z
C j
C j
Z
in
+ Φ
+ + =
ω
ω
, (1.8)

7
where Z
1
and Z
2
are the input impedances of the acoustic transmission line looking
towards the front acoustic port and back acoustic port, respectively. Equations that
describe insertion-loss and pulse-echo response et al can also be obtained with the KLM
model. From the design point of view, the KLM model allows an intuitive approach to
be used in optimizing the transducer performance: the two acoustic ports can be used to
interpret front matching and rear backing; the electrical port, on the other hand, can be
used to explain electrical matching. The popular transducer simulation software
PiezoCAD, which was extensively used in the project, is based on this model.
1.2.2 Mechanical Matching

Fig 1.3 A simplified model of an ultrasonic transducer.

As discussed previously, in medical ultrasound imaging applications, a transducer
with a broad bandwidth and low loss is preferred. Mechanical matching, which includes
front matching and rear backing, is a well-established solution to a high performance
medical transducer.
A simplified transducer model is illustrated in Fig 1.3. When a single electrical
pulse is applied across the piezo-electrical element, pressure waves a and b will be
produced with opposite directions on the front and the rear surfaces of the element,
respectively, as shown in the figure.
Z
p
Z
b
Z
l

b
a
Front Rear

8
The pressure wave b moves forward and reaches the front surface of the piezo-
element. The pressure transmits into the front loading medium at a normal incidence is
governed by the transmission coefficient:
2
l
p l
Z
T
Z Z
=
+
, (1.9)
where Z
l
and Z
p
are acoustic impedance of the loading medium and piezo-elements,
respectively. The acoustic impedance Z
p
of a typical piezo-element is normally around 30
MRayls, which is much higher than the acoustic impedance of the loading medium, such
as human tissues, which is around 2 MRayls. With the typical values, the transmission
coefficient T is calculated to be only around 10%. Therefore, a matching layer was
inserted in between piezoelectric materials and the loading medium to compensate for
their acoustic impedance mismatch. According to transmission line theory, 100 %
transmission occurs when thickness of the matching material is equal to λ
m
/4 ( λ
m
is the
wavelength in the matching material) and acoustic impedance of the matching material
Z
m
satisfies
2 / 1
) (
l p m
Z Z Z = , (1.10)
Desilets et al. (1978) found, however, for broadband transducers application, the
above equation should be modified to
3 / 1 2
) (
l p m
Z Z Z = , (1.11)
Sometimes, when the acoustic impedances of the piezoelectric materials are very
high, two quarter-wavelength matching layers are suggested. Desilets et al showed that

9
the acoustic impedances of the two matching layers should be respectively equal to
7 / 1 3 4
1
) (
l p m
Z Z Z = , (1.12)
6 1/7
2
( )
m p l
Z Z Z = , (1.13)
Popular matching materials include silver epoxy, glass, parylene et al.
Similarly, when the pressure wave a moves backward and hits the rear surface of
the piezo-element at a normal incidence, a fraction of energy will be transmitted into the
air and the rest is reflected back and is governed by the reflection coefficient:
a p
a p
Z Z
R
Z Z
−
=
+
, (1.14)
where Z
a
is the acoustic impedance of the air, which is close to zero. It is obvious that a
majority of the energy will be reflected back to the front face. In reality, even with
matching layers, 100% transmission is impossible at the front surface. The reflected wave
thus will reverberate inside the piezo-element, causing a long ring (narrow bandwidth) of
the produced ultrasonic pulse. Backing, therefore, is used to damp out the ringing due to
acoustic impedance mismatch between the air and piezoelectric materials. Ideally, when
the acoustic impedance of the backing (Z
b
) is equal to the piezo-element (Z
p
), there will
be no reflection (R = 0) and a monocycle pulse can be generated. However, the sensitivity
will be significantly reduced in this case. A compromise has to be made between
bandwidth and sensitivity in practical applications. In most of medical transducer
applications, conductive epoxy, E-solder 3022 (Von Roll Isola Inc., New Haven, CT),
was used. The material has a high attenuation (120 dB/mm at 30 MHz) and relatively low

10
acoustic impedance (5.9 MRayls), making it possible to achieve low insertion loss, short
and well-shaped pulses suitable for imaging purpose. Tungsten loaded epoxy is another
popular choice. It has higher acoustic impedance (8-20 MRayls) and is used when the
broad bandwidth is the priority consideration. Besides lossy, the backing materials are
usually rigid so as to support the fragile piezo-elements.

1.2.3 Electrical Matching
Transducers are often electrically tuned to maximize energy transmission
(decrease insertion loss) and improve their bandwidth. Ideally, for maximal energy
transmission, the transducer input impedance should be real (tune out the clamped
capacitance) and the input resistance should match the electric impedance of the source
(normally 50 Ω). The capacitance can be tuned out by simply add an inductance either in
series or in parallel with the transducer. It’s calculated that at series resonance (
r
f ) a
shunt inductor with value of 1/(ω
r
2
C
0
) (C
0
= ε
S
A/L , where A and L are the surface area
and thickness of the transducer element, respectively) may be used, Fig. 1.4 (a). For the
parallel resonance (
a
f ) case, a shunt inductor with value of 1/(ω
a
2
C
0
)+R
a
2
C
0
(where R
a
=
4k
t
2
Z
c
/πω
a
C
0
(Z
1
+Z
2
), Z
1
is acoustic impedance of load medium, Z
2
is acoustic impedance
of the backing material and
a
ω

is parallel resonance frequency) can be used, Fig.1.4 (b)
(Shung et al, 1996). To match the resistance to source impedance, a transformer may be
used. Hunt et al (1983) suggested that the closer the transducer’s resistance is to 50 Ω
before transformer tuning, the better the transducer’s bandwidth is after transformer
tuning.

11

(a) (b)
Fig 1.4 Serious (a) and parallel (b) equivalent circuit model for a single-element
transducer at its resonance.

1.2.4 Single-element Transducers
The most elementary ultrasound transducer is the single element transducer. A
broadband single element transducer normally consists of a piezo-element which was
sputtered with electrodes on both surfaces, matching layers, backing material and
sometimes a lens. The functions of the backing and matching layers have been explained
in the previous section. The lens here is used to focus the ultrasonic beam to a desired
distance. For a piston transducer with radius of a and wavelength of λ in the loading
medium, there is a distance which is the last maximum of the axial pressure. The distance
was found to be
λ
2
0
a
Z = , (1.15)
The region between transducer surface and this distance is called near-field zone
or Fresnel zone. In this zone, the axial pressure oscillates. The region beyond the distance
is called far-field zone, where the axial pressure decreases gradually. The far-field region
is the ultrasound imaging region. This means the target of interest should be always
located beyond Z
0.
R
a

C
0

R
r
C
0

12
In the far-field, directivity function of a piston transducer is found to be
1
2 ( sin )
( )
sin
J ka
H
ka
φ
φ
φ
= , (1.16)
where J
1
is the first order of the Bessel function. With the equation, the main lobe shape
can be related to
a
λ
ϕ 61 . 0 sin = , (1.17)
For a rectangular element which is the basic unit of an array, its directivity
function is
sin
sin ]
sin[ sin[ ]
2 2
( , )
sin sin
2
2
y
x
x y
x y
kb
kc
H
kc kb
φ
φ
φ φ
φ φ
= , (1.18)
And its main lobe angel can be related to
c
x
λ
ϕ
1
sin
−
= ,
b
y
λ
ϕ
1
sin
−
= , (1.19)
where c and b is the dimension c in x-direction and dimension b in y-direction
respectively.

1.2.5 Linear Arrays
The linear array consists of a large number of identical transducer elements of
rectangular cross section in a linear arrangement. Like a single-element transducer, a
backing material and one or two matching layers are used to improve its performance. To

13
minimize acoustic cross-talk, piezoelectric material and matching layers (sometimes
backing and the lens are even included) are diced into small elements. The space between
two elements (g) is called a kerf. The kerfs may be filled with acoustic isolating material.
The isolating material serves a dual purpose: it can decrease acoustic cross-talk, but it
also provides a rigid support for array elements. Sometimes, a lens is used. It also
provides a dual purpose. it can focus the acoustic beam in the elevation direction thus to
decrease slice thickness of the imaging plane which can cause serious image artifacts; at
the same time, it can serve as a protecting layer for the fragile array elements.
The radiation pattern of a linear array in the far field is given by
1
( ) sinc( ) ( ) sinc( )
N
m
bu Lu
H u u m
g
λ
δ
λ λ
=
= • − ∗
∑
, (1.20)
where u = sinφ
x
, H(u) is the direction function at an angle of φ
x
, L is length of the array,
N is the number of the elements, b is the width of the element, g is pitch which is the
space between the center of two adjacent elements, λ is the wavelength in the loading
medium.
The linear array radiation pattern equation indicates that at certain angles big side
lobes called grating lobes may occur. The angle where the grating lobe appears is
governed by the following equation.
) ( sin
1
g
n
g
λ
−
= Φ , (1.21)
where n is an integer. Therefore, the pitch should be less than λ to make sure the grating
lobes occur at angles larger than 90° (in a practical design, g is normally equal to 0.75 - 2

14
λ). With the above two equations, other rules for designing a linear array can also be
obtained: the width of the element should be as large as possible in order to damping the
magnitude of the grating lobes; the aperture of the array should be as large as possible in
order to get a narrower main beam width; and ratio of the width to thickness of the
element ( b/t ) should be < 0.6 to avoid lateral resonance. However, above requirements
can’t be satisfied at the same time, some trade-offs have to be made in a practical design.

1.2.6 Issues in High Frequency Ultrasonic Imaging
High frequency ultrasonic imaging provides improved spatial resolution. This
benefit makes it a promising imaging tool in clinical applications and academic research.
But there are several issues which limit its applications and manufacture. A major
drawback of high frequency ultrasonic imaging is its penetration depth. High frequency
ultrasonic imaging also brings challenges in system design and transducers fabrication.
Penetration issue is caused by attenuation of ultrasonic energy in the loading
medium. The energy of ultrasonic waves will be attenuated, due to absorption and
scattering, as a function of traveling distance when waves propagate into a load medium.
The attenuation can be mathematically described as
0
z
P P e
α −
= , (1.22)
where P
0
is the ultrasonic pressure at starting point of propagation, α is attenuation
coefficient and z is the traveling distance. Attenuation coefficient α is frequency
dependent and also has strong dependence on the type of tissues. It can be defined as
γ
α α f
0
= , (1.23)

15
where
0
α is the attenuation coefficient at 1 MHz and γ is the frequency dependence
parameter. Table.1.1 shows attenuation coefficients of a variety of tissues from 1 to 10
MHz. Attenuation coefficient at higher frequency (10-100 MHz) was summarized and
plot in Fig 1.5. To illustrate the importance of the working frequency on the ultrasound
penetration, live is selected here as an example. At 5 MHz, its attenuation coefficient is
4.5 dB/cm. For an imaging system with dynamic range of 80 dB, the penetration of up to
9 cm is possible. However, at 100 MHz, as shown in the Fig 1.5, the attenuation
coefficient of live reaches 14 dB/mm (140 dB/cm). With the same dynamic range, the
penetration is only 3 mm. So applications of high frequency ultrasonic imaging are those
which require better spatial resolution but less penetration.
Table 1.1 Frequency dependence of α applies at the range 1-10 MHz of some
tissues relevant to ultrasonic imaging. Data collected by Duck (1990) and Wells (1999).
Material Attenuation coefficient α at 1 MH
(dB/cm)

Frequency dependence of α
Air 1.2
2
f

Blood

0.2
1.3
f

Fat 0.6 f
Liver 0.9 f
Soft tissue 0.6 f
Water 0.002
2
f

16

Fig 1.5 Attenuation of variety tissues in the frequency range from 10 to 100 MHz, from
Foster (2000)

For a thickness mode transducer, its center frequency is inversely proportion to
the thickness of the piezoelectric elements. For those transducers working in the
thickness mode, the thickness is inverse proportional to the center frequency; the
precision of lapping and dicing will be critical issue at high frequency. Though, PZT
ultrasonic arrays up to 35 MHz have been reported by Cannata et al (2006) using
conventional diced and fill techniques. Fabrication of even higher frequency transducers
(>50 MHz) has bottlenecked. For a 100 MHz transducer, even with high sound speed
materials such as lithium niobate, a thickness of less than 20 µm is required.
Consequently, the other dimensions such as aperture size and kerf of the transducer will
also have to be scale accordingly. Even with precise dicing, it’ll be almost impossible to
fabricate it. Meanwhile, grain sizes of the bulk piezoelectric materials (normally 2~5 µm)
are too large for very high frequency applications. During recent years, the study of

17
the Micro Electro Mechanical System (MEMS) has offered significant opportunities for
miniaturized devices. At the same time, PZT film has been made much progress in
piezoelectric properties and shows a number of advantages in making high-frequency
ultrasound transducers and arrays. So integrating PZT films into MEMS devices may be a
feasible solution to high- frequency transducers and arrays fabrication.
1.3 Scopes of The Research
The goal of this research is to demonstrate the PZT thick films is a promising
piezoelectric material for high frequency applications, and integrating PZT thick film into
MEMS techniques is a feasible approach for fabrication of high frequency transducers.
This thesis describes the design, fabrication and characterization of very high frequency
(~100 MHz) single-element transducers, 32-element kerfless linear arrays and 32-element
kerfed linear arrays with PZT thick films.
Chapter 2 provides the background knowledge for piezoelectric materials.
Important ferroelectric and piezoelectric properties of the piezoelectric materials are
briefly explained. Different piezoelectric materials are reviewed and compared.
Especially, their applications in high frequency ultrasound transducers are discussed.
Chapter 3 describes production of PZT thick films. Advantages of the PZT films to other
piezoelectric materials are presented; fabrication process is described; approaches for
improvement of film qualities are investigated; physical properties of the PZT films are
measured. The modeling and fabrication of the PZT thick film single-element transducer
and the kerfless array are included in Chapter 4. The performance of the PZT film
kerfless array is characterized and compared with a PZT sheet kerfless array fabricated

18
with the same array pattern. At the end of this chapter, the acquisition of wire phantom
imaging, porcine eyeball imaging and human skin imaging using PZT thick film single-
element transducer are presented. In Chapter 5, we demonstrate the development of a
PZT film kerfed array by MEMS dry-etching technique. In Chapter 6, the results
presented in the research are summarized and the future work is suggested.

19
CHAPTER 2
INTRODUCTION TO PIEZOELECTRIC MATERIALS

Transducers for ultrasonic imaging are almost always fabricated from
piezoelectric materials. Important properties of the piezoelectric materials are first
introduced. Different piezoelectric materials include piezoelectric ceramics,
polyvinylidene fluoride (PVDF), relaxor-based single crystals, composites and
piezoelectric films are reviewed and compared at the final sections. Especially, their
applications in high frequency medical transducers are discussed.

2.1 Introduction
Piezoelectric materials are a class of materials which can be polarized, in addition
to an external electric field, also by application of a mechanical stress. One of the most
important properties of the piezoelectric materials is piezoelectric and inverse
piezoelectric effects. Piezoelectric effect may be written as
i ijk jk
D d X = , (2.1)
where D
i
is charge density, X
jk
is applied stress and d
ijk
(CN
-1
) is piezoelectric coefficient.
It states that an electrical potential appears across the material surfaces by application of a
mechanical stress on it.
Inverse piezoelectric effect may be expressed as
t
ij kij k ijk k
x d E d E = = , (2.2)

20
where x
ij
is the strain, E
k
is the external electric filed and d
kij
(mV
-1
) is converse
piezoelectric coefficient. It describes change in dimension of the piezoelectric materials
when an electric field is applied. The piezoelectric coefficients d for direct and converse
piezoelectric effects are thermodynamically identical.
The two effects are the basic of the ultrasound transducer which performs the
conversion of electrical energy into mechanical energy (inverse piezoelectric effect), and
conversely, the conversion of mechanical energy into electrical energy (direct
piezoelectric effect).
Ferroelectric materials are the piezoelectric materials whose direction of
spontaneous polarization, which is defined by the value of the dipole moment per unit
volume, can be switched by an external electric field. Ferroelectric properties of a PZT
film can be used to assess quality of the film and even to predict potential performance of
the PZT film transducers. Therefore, in the chapter, we’ll first introduce several basic
ferroelectric properties of piezoelectric materials which include phase transition, poling
mechanism, and ferroelectric hysteresis loop; morphotropic phase boundary is then
explained with the PZT as the example. Their correlations with PZT film are emphasized.
At the end, popular piezoelectric materials are introduced and compared with each other,
their high frequency application are discussed.

21
2.1.1 Phase Transitions and Perovskite Structure
An important characteristic of ferroelectrics is the temperature of phase transition
called the Curie point Θ
c
. When the temperature decreases through the Curie point, a
ferroelectric crystal undergoes a structure phase transition from a paraelectric phase, in
which electric dipoles are unaligned and has the potential to align in an external field, to a
ferroelectric phase, in which the spontaneous polarization can be reversed by an applied
electric field, and the crystal shows ferroelectricity. When the temperature is above Curie
point, the ferroelectric crystal does not exhibit ferroelectricity. It is generally believed
that the ferroelectric structure of a crystal is created by a small distortion of the
paraelectric structure such that the lattice symmetry in the ferroelectric phase is always
lower than that in the paraelectric phase.
Perovskite is the name of the mineral calcium titanate (CaTiO
3
). Most of
ferroelectric ceramics have pervoskite-type structure. Those ceramics have the general
chemical formula ABO
3
, where A represents a cation with a larger ionic radius, B
represents a cation of a smaller ionic radius, and O is oxygen. Other perovskite-type
materials have similar unit cells. The unit cell has a paraelectric phase with cubic
symmetry when the temperature is above its Curie point. At this situation, there is no
spontaneous polarization with the unit. When the temperature is below the Curie point,
the tetragonal ferroelectric phase occurs. Spontaneous polarization along the C
T
-axial
direction can be observed because of a small distortion of the paraelectric structure.

22
2.1.2 PZT Solid Solution and Morphotropic Phase Boundary
When Ti ions in PbTiO
3
are partially replaced by Zr with a molar ratio x, a solid
solution of xPbZrO
3
-(1-x)PbTiO
3
(0 < x < 100) binary system is formed. This solid
solution is called lead zirconate titanate (abbreviated as PZT) and its chemical formula is
Pb(Zr
x
Ti
1-x
)O
3
. PZT has the perovskite structure similar to that of PbTiO
3
. PZT is after
quartz the most widely used piezoelectric.

Fig 2.1 Phase diagram of Pb(Zr, Ti)O
3
solid solution, from Y.H Xu (1991)

Preparation in ceramic or film form has the advantages that the composition may
be more easily adjusted than when preparing single crystals, so that materials with a very
large range of properties may be obtained. A schematic of lead zirconate titanate T-x
phase diagram is shown in Fig 2.1. The structure of the high-temperature paraelectric
phase is cubic. Titanium-rich compositions transform into tetragonal perovskite structures
whereas the phase transformation in Zr-rich compositions is more complex. At low Ti
content (x > 95 %) and at room temperature the structure is orthorhombic and for x < 90

23
% the structure is rhombohedral, with another phase transition from a high to a low-
temperature rhombohedral phase. The rhombohedral and tetragonal structures in the
middle of the diagram are ferroelectric. From the application point of view the most
interesting compositions lie near the centre of the phase diagram, where tetragonal and
rhombohedral phases are separated by a boundary which is nearly independent of
temperature. The composition of this so-called morphotropic phase boundary (MPB) is
approximately 52/48 at room temperature.
The piezoelectric coefficients, electromechanical coupling coefficients, dielectric
permittivity and remanent polarization measured on ceramic samples reach a peak in the
region of the morphotropic phase boundary. Properties of random-oriented
polycrystalline PZT films are in qualitative agreement with results obtained from bulk
ceramics. Therefore, in this research, PZT sol-gel was always prepared around the ratio
of the 52/48 in the development of the PZT thick films. But a 20 mole% excess PbO was
added to the sol-gel solution to compensate for lead loss during heat treatments,

24
2.1.3 Poling Mechanism of Ferroelectrics

Fig 2.2 Ferroelectric with random orientation of grains before and after poling,
from Y.H Xu (1991)

Due to the complex set of elastic and electric boundary conditions, ferroelectric
grains in ceramics and polycrystalline films are always split into many domains. If the
direction of the spontaneous polarization through the material is randomly distributed, the
materials will show zero net polarization. Ferroelectric materials may be transferred into
a polar state by applying a strong electric field at elevated temperatures as shown in Fig
2.2. This process, called poling. Poling cannot orient grains, but can reorient domains
within individual grains approaching to the direction of the electric field. A poled
ferroelectric exhibits piezoelectric properties. Grown ferroelectric single crystals usually
contain many domains and may exhibit weak piezoelectric properties. The single-domain
state (a single crystal does not contain domains) in single crystals may be achieved by
poling. Non-ferroelectric material with randomly oriented grains cannot be poled and
exhibit piezoelectric properties. The polarization after the removal of the field (at zero
field) is called remanent polarization, P
r
. Maximum remanent polarization that may be
achieved in a material depends on available domain states. The actual polarization is in
fact always lower, as many domains cannot be reoriented and some domains will switch

25
back after the poling field is removed.
Poling conditions of bulk piezoelectric materials are not suitable for the
piezoelectric films. For example, the poling electric field of PZT films should be much
higher than that of PZT ceramics (normally 2~ 3 volts/µm) (Damjanovic, 1998). This is
because that films exhibit much higher coercive fields (typically 50 kV/cm to 100 kV/cm)
and higher breakdown voltages (200 to 400 kV/cm). Therefore, it is necessary to drive
films with higher fields in order to compensate partially for the smaller thickness. The
poling conditions of the PZT thick films will be investigated in the following chapter.

2.1.4 Ferroelectric Hysteresis Loop

Fig 2.3 Ferroelectric (P-E) hysteresis loop, from D. Damjanovic (1998)

The most important characteristic of ferroelectric materials is polarization reversal
by an electric field. This characteristic can be visualized with the ferroelectric hysteresis
loop. The hysteresis loop can be observed experimentally by using a Sawyer–Tower

26
circuit. Fig 2.3 shows a standard plot of a hysteresis loop. At small values of the AC
electric field, the polarization increases linearly with the field amplitude. This
corresponds to segment AB in the figure. In this region, the field is not strong enough to
switch domains to the direction of polarization. As the field is increased, the polarization
of domains with an unfavorable direction of polarization will start to switch to the
direction of the field (segment BC). The polarization response in this region is strongly
nonlinear. Once all the domains are aligned (point C) the ferroelectricity again behaves
linearly (segment CD). If the field strength starts to decrease, some domains will back-
switch, but at zero field the polarization is nonzero (point E). To reach a zero polarization
state the field must be reversed (point F). Further increase of the field in the negative
direction will cause a new alignment of dipoles and saturation (point G). The field
strength is then reduced to zero and reversed to complete the cycle. The value of
polarization at zero field (point E) is called the remanent polarization P
r
. The field need
to bring the polarization to zero is called the coercive field, E
C
. The spontaneous
polarization P
S
is usually taken as the intercept of the polarization axis with the
extrapolated linear segment CD.
There are important differences in the hysteresis loops measured on bulk
piezoelectric materials and piezoelectric films with similar composition. The most
important difference is the magnitude of the coercive field E
C
which may be as much as
ten times higher in films than in bulk materials of the same composition. The origins of
this difference are presently not clear. In addition, loops in films are often more tilted and
remanent polarization generally lower than in corresponding bulk materials. Some
difference may be attributed to different processing of films and bulk materials. It is

27
also speculated that the thickness, grain size and porosity of films and bulk materials play
an important role for these differences.

2.2 Piezoelectric Materials
There are a wide range of piezoelectric materials available to produce ultrasonic
waves. But only some of them are suitable for medical ultrasound imaging. In this section,
we will compare several popular piezoelectric materials: piezoelectric ceramic, relaxed-
based crystals, PVDF and its copolymers, composites, and piezoelectric films. The
general properties of these materials will be reviewed and summarized. Their applications,
especially in high frequency medical transducers, will be discussed.
A. Piezoelectric Ceramics
The Piezoelectric ceramics have very high electromechanical coupling, broad
range of dielectric constant and very low loss. Those properties have made them the most
popular materials for medical ultrasonic transducers in past decades. The piezoelectric
ceramics can be created by mixing piezoelectric powder with a binder, followed by high-
pressure pressing and high-temperature firing. The ceramic, overall, is approximately
isotropic because the domains are randomly orientated. A proper poling process has to be
performed to enhance its piezoelectric properties: apply a DC electric field at a
temperature close to Curie point (100-200 °C), and then the temperature is slowly
decreased with the electric field present. The barium titanate (BaTiO
3
) and lead zirconate
titanate (PZT) are the two most popular piezoelectric ceramics in the medical ultrasound
application. Especially, PZT ceramics have strong piezoelectric properties, making them
the most popular piezoelectric materials for medical transducer.

28
We can obtain a wide range of PZT materials optimized for different applications
by small composition change. Among them, PZT5A and PZT5H are suitable for medical
imaging application, but unsuitable for power application because of their relatively high
losses. PZT4, on the other hand, can be used in power application such as HIFU (High
Intensity Focused Ultrasound) because it has low loss and high Curie temperature (~300
°C). However, these conventional PZT ceramics is not suitable for building high
frequency transducers: they have grain size of 3-5 µm. In recent years, fine grain PZT,
such as TRS600FGHD (~ 1-µm grain size), has been developed for high frequency
applications.
However, all of the above ceramics have a main drawback: their acoustic
impedances are very high (~30MRayls). Acoustic matching layers thus have to be
fabricated to overcome the huge acoustic mismatch.
B. Relaxor-based Crystals
The lead magnesium niobate-lead titanate (PMN-PT) and lead zinc niobate-lead
titanate (PZN-PT) are new generation of relaxor-based ferroelectric single crystals. Their
exhibit electromechanical coupling coefficients of greater than 0.9, which is a major
improvement considering that the highest value for PZT ceramics is around 0.7. Their
superior properties make them very useful for making high-sensitivity broad bandwidth
medical transducers. Moreover, unlike piezoelectric ceramics, the crystals don’t have
grain size consideration, thus can be used in high frequency applications. But they share
the same drawback with piezoelectric ceramics: high acoustic impedance (>30 MRayls).
In addition, they are not suitable for high temperate applications because of their
relatively low curie temperatures ( ~ 150 ºC).

29
C. PVDF and Its Copolymers
Piezoelectric polymers such as Polyvinylidence difluoride (PVDF) are another
kind of popular materials for medical ultrasonic transducers. The piezoelectric property
of PVDF was firstly discovered by Kawai in 1969 by stretching the film and applying an
electric field of 300 kV/cm at a temperature of around 100 °C. PVDF has built its
reputation as a useful transducer material because of its many advantages over
piezoelectric ceramics and crystals: its acoustic impedances (~ 4 MRayls) are close to
human tissues (~ 2 MRayls), broad bandwidth can be obtained even without acoustic
matching; it is very flexible, allowing press-focused. Typical thickness of the
commercial available PVDF film ranges from several microns to around twenty microns,
thus it is mainly used for building high frequency applications. Particularly, because of its
close impedance match to water and high receiving constant, it is an excellent material
for high frequency ultrasound receiver, such as hydrophone. But its low
electromechanical coupling, very low dielectric constants, high dielectric loss have
limited their usages in many applications, although recent development of P(VDF-TrFE)
co-polymers has shown a higher electromechanical coupling coefficient.
D. Piezoelectric Composites
Development of piezoelectric composites has attracted lots of interest because
they have almost all the properties needed for building high-sensitivity, broad-bandwidth
transducers used in medical ultrasonic imaging. Their electromechanical coupling can be
even higher than most popular used piezoelectric materials. At the same time, the
composites have low acoustic impedance (10-20 MRayls), broad range of dielectric
constants, low dielectric and mechanical loss. Besides, the composite materials are soft,

30
so press-focused method can be easily used in those materials. Among the composite
structures, 1-3 composite is the most popular one. We will investigate its feasibility of the
high frequency application in the following paragraphs.
In a 1-3 composite transducer element, its first lateral mode frequency has to be at
least two times larger than its operating frequency, which can be stated as:
2
2 2 2
f
c
V
V
t w
≥ ×
× ×
, （2.3）
where V
f
and V
c
is the sound velocity of the filler and the composite material,
respectively, w is the width of the kerf to be filled with epoxy and t is the thickness of
the finished composite material. With the equation (2.3), the profile angle, which is
defined as the ratio of element depth (t) to a half of the kerf width (w/2), has to meet:
4 2
tan( )
2
c
f
V t
w
V
α
×
= ≥ , (2.4)
In low frequency applications, the composite can be fabricated by dicing-and-
filling method. But in high-frequency applications, the size of the kerf required is very
difficult to be achieved by the dicing saw. Assume we want to build a 40 MHz transducer
with PMN-PT composite. The sound velocity of the PMN-PT composite (V
c
) and epoxy
filler (V
f
) is normally around 3200 m/s and 1100 m/s, respectively. Substitute these data
into the equation (2), we obtain tan(α) ≥ 16. This requires the profile angle, α, to be no
smaller than 87º. Consequently, the filling kerf of the composite structure cannot be
wider than 5 µm to suppress the effect of the lateral mode resonance, which is very
difficult to be cut with the conventional dicing saw. The MEMS micromachining
technology or laser dicing are possible solutions to the problem (X. Jiang, et al).

31
However, if we want to construct a composite transducer operating at a center frequency
higher than 100 MHz, the kerf of the composite structure will be only 2 µm. Even state-
of-art MEMS technique or laser dicing is not a practical approach to producing such
narrow kerf.
E. Piezoelectric Films
Piezoelectric films are a promising solution for miniaturized sensors, actuators,
filters, and high- frequency ultrasonic transducers. Among the piezoelectric films
materials, Aluminum nitride (AIN), Zinc oxide (ZnO), Lead Zirconate Titanate (PZT) are
most popular used because of their extraordinary piezoelectric properties, which are
desired in MEMS devices. Important properties of the piezoelectric films are listed in
Table.2.1.
Zinc Oxide (ZnO) is the first piezoelectric material used for commercial thin-film
applications. It has the strongest piezoelectric effect among non-ferroelectric materials.
Strong piezoelectric effect, great stability and availability make it one of the most popular
piezoelectric thin films.
Aluminum nitride (AIN) is another popular thin-film material due to its high
frequency constant (high acoustic velocity), endurance in high temperature and humidity
and it’s mechanically sturdy. Both ZnO and AIN are wurzite materials and are non-
ferroelectric piezoelectric. They show a piezoelectric response along [001]. Most of the
ZnO and AIN thin films are sputter deposited As shown in Table 2.1, AIN exhibits low
loss (0.003), low dielectric constant (10.1), high acoustic velocity (10400 m/s) and
piezoelectric coefficients are moderate (3.9 pm/V). ZnO and AIN thin films exhibit very
similar piezoelectric properties. The transverse coefficient is almost equal, while the

32
longitudinal coefficient and coupling coefficient of ZnO are larger AIN’s.
PZT film is a perovskite material and is ferroelectric. It possesses high dielectric
constant, low velocity and large coupling coefficient. From the standpoint of transducer,
PZT film has a clear advantage due to its good piezoelectric and dielectric properties.
Sol-gel deposition method has been widely used for producing PZT thin films. One
drawback of the PZT film however is its piezoelectric properties are strongly influenced
by orientation, composition, grain size, defect chemistry, and boundary conditions. Thus
it is very difficult to deposit high quality PZT thin films repeatedly. It is the main task of
this research to resolve this problem. In the next chapter, we will present how to produce
high quality PZT film in detail.
Table 2.1. Summary of the ZnO, AIN and PZT films piezoelectric and dielectric
properties
ZnO AIN PZT
Mechanical loss (at 1KHz)
0.01-0.1 0.003 0.01-0.03
) / (
2
, 31
m C e
f

-1.2 -1.05 -9.6
) / (
. 33
V pm d
f

7.5 3.9 144
Acoustic velocity ( m/s) 6330 10400 4000
f , 33
ε

9.5 10.1 650
Voltage response
33 0 . 31
/ ε ε
f
e

) / ( M GV

-7.2 -10.3 -1.4
Coupling coefficient on silicon
2
.
) (
f p
k

0.06 0.11 0.19

33
Summary
The important properties of different piezoelectric materials are listed in Table 2.2
(Xu et al., 2002). Among them, PVDF has the lowest dielectric constant, coupling
coefficient and acoustic impedance. It can be used in high –frequency broadband
applications, but is not suitable for miniature size devices, such as array element, since its
dielectric constant is extremely low. Both PZT5H and PMN-PT have high coupling
coefficient and high dielectric constants. They are still important materials for medical
transducers, even though their acoustic impedances are high. Fine-grain PZT and PMN-
PT single crystal can be fabricated into high-frequency transducers as long as they don’t
crack during the mechanical lapping process. PMN-PT composite is a promising
piezoelectric materiel. It exhibits excellent properties in medical transducers applications.
But fabrication cost is a big concern. Besides, there is still not a practical solution to very
high frequency (>100 MHz) transducers. PZT film, which is the subject of the research,
shows medium-range values among the piezoelectric materials. In the following chapter,
we will study the feasibility of fabricating very high-frequency (100 MHz) medical
transducers with the materials. Its advantages and drawbacks will be discussed and
compared with the conventional piezoelectric materials.

34
Table 2.2 Properties of different piezoelectric materials
PVDF PZT5H PMN-PT PMN-PT 1-3
composite
PZT film*
Dielectric constant
(at 1 KHz)
10 3400 ~3000 ~800 1000-2000
k
t
0.11 0.55 0.6 0.8 0.2-0.35
Acoustic
impedance
(MRayl)
3.4 34 37 15 15-20

*the PZT film data varies with fabrication conditions.

35
CHAPTER 3
PZT THICK FILM FABRICATION
Piezoelectric thick-films were first described by Baudry in 1987, who fabricated
an acoustic coupler to demonstrate an application of the technology. It represents a
structure which consists of substrate, bottom electrode, piezoelectric layer and top
electrode. Ferroelectric compositions are almost exclusively based on lead-containing
perovskites like lead zirconate titanate (PZT) or complex perovskites represented by
Pb(Mg
1/3
Nb
2/3
)O
3.
PZT thick film normally has a thickness of 10-30 µm. This thickness
range makes the films possible to be fabricated into high-frequency (>50 MHz) ultrasonic
transducers.

3.1 Introduction to PZT Thick Film
PZT thick-film offers a number of advantages in high-frequency ultrasonic
transducers applications. These include:
(1) PZT thick-films have relative lower acoustic impedance ( ～15 MRayls) compared
to popular bulk piezoelectric materials ( ～ 30 MRayls). This makes acoustic
impedance easier to be matched to loading mediums (normally ~2 MRayls). As a
result, PZT film transducers offer broad bandwidths.
(2) PZT films have submicron grain sizes. To build transducers operate at 100 MHz or
even higher, the grain size of the piezoelectric materials should be less than one
micron, otherwise, the performances of the transducers will deteriorate. While
grain sizes of bulk piezoelectric materials normally are larger than 2 microns.

36
(3) During recent years, the study of the Micro Electro Mechanical System (MEMS)
has offered significant opportunities for miniaturized devices. Integrating PZT
films into MEMS devices maybe a feasible solution to high- frequency transducers
and arrays fabrication. MEMS uses traditional semiconductor technologies. This
makes bulk production PZT films possible. And from the point of view of
electronic circuit design, PZT films are CMOS compatible. The resulting signals
can be processed on chip.
The use of PZT films in sliver-mode high-frequency ultrasonic transducers
applications requires thick, dense and crack-free films with excellent piezoelectric and
dielectric properties. The key technique of the PZT thick film ultrasonic array fabrication
is how to prepare high quality PZT thick films, which will ultimately determine the
performance of the ultrasonic transducers built with them. While it’s difficult to obtain
crack-free PZT films thicker than 7 µm solely using the PZT sol-gel solution (Zhou et al
(2005)). Instead, PZT powder/so-gel solution composite was suggested to deposit PZT
thick films. This is because adding PZT powder to PZT so-gel solution can reduce the
stress within the films significantly; thus it is possible to obtain films thicker than 10 µm
without crack.
There are a number of methods for making thick-films. Among them, the
following four methods are more popular.
Screen printing: A ceramic paste is deposited onto the substrate through a mesh
using a rubber blade which is allowed to dry and then sintered. The mesh thickness
determines the final film thickness and the masked areas determine the ceramic pattern
deposited on the surface. This process makes it possible to produce films from a few to

37
several tens of microns. A major disadvantage of this method is the inability to make
micrometer-sized features.
Tape-casting: Ceramic slurry is spread evenly onto a flat horizontal surface.
Once dry, the sample is cut, laminated or shaped and sintered.
Aerosol deposition: Submicron particles form an aerosol flow by mixing with a
carrier gas. This flow is accelerated an ejected from a nozzle into a deposition chamber
where the particle bombard the substrate to form a thick film.
However, it is difficult to process crack-free PZT films thicker than 10 µm on Si
substrates using either chemical solution deposition (Dey et al., 1988; Kurchania et al.,
1999) or most vapor deposition methods (Sakashita et al., 1993). Techniques such as
screen-printing (Gourverneur et al., 1993; Mass et al.,1997; Zhu et al., 2000) , tape-
casting (Biggers et al., 1979; Navarro et al.,2004), and aerosol deposition (Akedo et al.,
2000; Maeda et al., 1998) have been used to fabricate thick films (> 15 µm). Screen
printing, tape-casting and aerosol deposition methods produce PZT films of low density
as a result of the low stress applied to the powder particles during shaping. The composite
spin coating method however can fabricate high density PZT films; thus it was selected
for this work.
Composite spin coating: The composite ceramic sol-gel film technique has been
proved to be a successful technique to fabricate thick PZT films on various substrates in
which, a chemical sol gel is generally loaded with appropriate concentrations of ceramic
powers. The resultant slurry type composite sol-gel will be spun onto the substrates
followed by optimized pyrolysis and annealing steps. This approach enables to fabricate
controllable, crack-free PZT thick films (> 10 µm) with minimized stress levels

38
between various layers. These advantages of composite ceramic sol-gel technique are
due to the formation of a strongly bonded network between the sol gel and ceramic
particles along with enhanced adhesion with the substrates minimize the cracks in the
resultant thick films. The success of this process was summarized initially by Barrow et
al., in ferroelectric PZT thick films having piezoelectric properties (d
33
= ~ 325 pC/N and
d
31
= ~ -80 pC/N) comparable with those of the bulk PZT ceramics.

3.2 Fabrication of PZT Thick Film
High quality PZT films cannot be grown directly on silicon. Buffer layers are
needed to prevent interfusion and oxidation reactions of PZT film with silicon.
PZT/Pt/Ti/SiO
2
/Si is the most widely applied sequence. Ti is an adhesion layer. Platinum
(P
t
) does not inhibit the diffusion of Ti to PZT side, where it reacts with oxygen and
servers as nucleation centers for PZT. As the lattice constant of the P
t
is rather close to
that of PZT, the PZT tend to grow with (111) direction on P
t
. Platinum coated silicon
wafer purchased from Nova Electric Materials, Ltd. was used as the substrate in this
project.
A 2-methoxyethanol-based sol-gel Pb(Zr
0.52
Ti
0.48
)O
3
(PZT) precursor was used as
the matrix. To compensate for lead loss during heat treatments, a 20 mole% excess PbO
was added to the solution adjusting its concentration to 0.3 M. A PZT 5H powders were
procured from Piezoelectric Technology Inc., to use as the loading powder in the
composite sol-gel. However, the as-received powder has an average particle size of ~5
μm and hence these ceramic powders were ball milled initially in an ethanol medium

39
using a Fritsch Pulverisette milling machine. After ball milling, the resultant powders
were found to have an average particle size of ~ 400 nm. Subsequently, these powders
were mixed with PZT sol-gel precursor and subjected to further ball-milling to obtain
well-dispersed composite solution.
The prepared composite solutions were spun at 2000 rpm for 30 seconds on
platinum coated silicon substrates using a Chemat spinner. Each PZT layer was
subjected to a two-step pyrolysis scheme, one at 200
°
C for two minutes in air followed
by second step at 400
°
C for two minutes in air. Subsequently, each layer was annealed
at 750
°
C for one minute in a rapid thermal annealer. The above process was repeated
multiple times until the desired thickness was achieved. The whole fabrication process is
illustrated in Fig.3.1.
Reducing film porosity in composite spin-coating method is a serious issue to
obtain superior dielectric and piezoelectric properties in thick PZT films. The PZT sol-
gel solution infiltration into the PZT composite films could reduce partially the porosity
in thick PZT films (Dorey et al., 2002; Corker et al., 2002). Besides, the ball milling time
and sol-gel solution to loading powder mass ratio of PZT composite solutions also are
very important factors influencing film physical properties. In the following paragraphs,
we’ll conduct a number of experiments to develop a practical recipe with which high
quality PZT film could be produced.

40

Fig 3.1 Flowchart of the PZT film fabrication process

3.2.1 Sol-gel Infiltration
Generally, the thick films derived from the compound solution are porous because
particles are dispersed in the solution as shown in Fig 3.2 (a), which degrades the
dielectric and piezoelectric properties of the films. To obtain dense films, an important
procedure is that sol-gel thin layers and composite-derived thick layers are deposited
alternately. The details of the process are described as follows:
PZT solution (0.37 M) was dipped on the surface of the thick film and vacuumed
around 60 seconds to make sure that the thick film was well infiltrated with PZT sol-gel

PZT sol-gel PZT powder
Mixed and ball-milled
Spin-coating
Two-stage pyrolysis
Annealed at 700 ºC
Crystallized at 800 ºC

41
solution. The vacuumed sample was then spin-coated at 2000 r/m for 30 seconds.
Afterward, the film was sintered for two minutes at 200 °C and another two minutes at
400 °C followed by final sintering with RTA at 700 °C for one minutes. The films were
monitored using a scanning electron microscope (SEM, Hitachi, S-3500N, Tokyo, Japan).
With filling process, the films become denser as shown in Fig 3.2 (b), which in turn will
enhance the dielectric and ferroelectric properties of the films.

(a)

(b)
Fig 3.2 SEM pictures of a PZT thick film without (a) and with (b) sol-gel infiltration
process

42
3.2.2 Milling Time Control
To get a well-dispersed composite solution, the milling time of the steel ball is a
key parameter. Particle Size Distribution (PSD) analyzer (Particle Sizing Systems, Inc.
Santa Barbara, Calif., USA) (Particle Sizing Systems, Inc. Santa Barbara, Calif., USA)
was used to estimate the particle sizes of the ball-milled PZT powders. The composite
solution was divided into three groups. There were milled 10 hours, 20 hours and 30
hours respectively. The average particle size is 411 nm after milling 10 hours with
ethanol as shown in Fig 3.3 (a). However, the average diameter is about 250 nm after
PZT composite solution was milled 20 hours (see Fig3.3 (b)). When the composite
solution was milled for 30 hours, the average particle size decreased slightly to 220 nm as
shown in Fig3.3 (c). As to the stand deviation of the particle size, the 20-hour group
drops significantly to 195 nm from 249 nm (10-hour group). But 30-hour group shows
almost the same value as the 20-hour one. The results suggest that 20-hour is a proper
period of time to get a well-dispersed PZT composite solution.

43

(a)

(b)

(c)

Fig 3.3 Particle size distribution of the PZT composite solution after 10-hour (a), 20-hour
(b) and 30-hour milling (c).

44
3.2.3 Powder to Solution Mass Ratio
The powder to solution mass ratio has great effect on the quality of the thick films.
A correlation between the effect of ceramic powder loading into sol-gel on the
microstructure and electrical properties needs to be addressed to optimize viable
processing conditions for thick PZT films for various ultrasonic bio-medical imaging
applications. For this purpose, a set of PZT films were prepared following composite
ceramic sol-gel method with solution-to-powder mass ratios of 1, 2, 4 and 6.
Circular Cr/Au electrodes with a diameter of 1.5 mm were deposited by sputtering
as top electrodes onto the films for a quantitative comparison of the functional properties
of these films. Dielectric properties were measured using an Agilent 4294A impedance
analyzer and polarization - field (P-E) hysteresis properties were evaluated using Radiant
precision materials analyzer. Fig. 3.4 and Fig.3.5 show a comparison of the ferroelectric
hysteresis and dielectric properties of the films fabricated using composite solutions with
variable solution-to-powder mass ratio, respectively. In addition, the dielectric constant
and remanent polarization values as a function of solution-to-powder mass ratio are
shown in Fig. 3.6. The dielectric constant and remanent polarization values increased
with the increasing proportion of PZT solution. With a solution-to-powder mass ratio of
0.5, the film exhibited ~8 µC/cm
2
remanent polarization value and this increased to ~ 37
µC/cm
2
for a mass ratio of 6. Similarly, the dielectric constant increased from 450 to
1250 at 1 kHz (the dielectric loss is about 0.03 at 1 kHz), which is much higher than the
previously reported results (Luckas et al., 2000). These results are comparable with
earlier works (Corker et al., 2002; Dorey et al., 2002); and indicate that comparable

45
dielectric properties can be achieved with optimized solution-to-powder ratios combined
with sol-infiltration procedure with no sintering aids.

Fig 3.4 Hysteresis loop of the thick films with different solution-to-powder mass ratio

Fig 3.5 Dielectric properties of the thick films with different solution-to-powder mass
ratio

46

Fig 3.6 Dielectric constant and remanent polarization values as a function of solution-to-
powder mass ratio

The effective transverse piezoelectric coefficient (e
31,f
) of these films were
measured by a modified wafer flexure method (Shepard et al., 1998). In the method, the
charge developed on the capacitor is measured using the current value between the top
and bottom electrodes (with a lock in amplifier at 4 Hz frequency and 0.5 V

signal to
drive the speaker).
θ ω sin
out
i
C = , (3.1)
where C is the charge, ω is angular frequency and θ is phase angle in degrees. The strain
imposed on the piezoelectric film during the sinusoidal pressure excitation was measured
using a bridge technique with a strain sensor (from Omega). To minimize the error due
to the sample shape, the output voltage (V
out
) is measured in two orthogonal positions

47
using the strain gauge and the mean of the two readings is taken to calculate the strain on
the wafer. The formula to calculate the strain is shown below (with gauge factor):
0.001
0.001
2.11 (1 )
2
out
out
V
Strain
V
×
=
×
× +
, (3.2)
where
out
V = output voltage in mV. Finally, the effective e
31
values are measured using
the following formula:
31
arg
2
ch e
e
area strain
=
× ×
, (3.3)
The measured results were summarized in Fig. 3.7. As shown in the figure, the
e
31,f
values increased with the solution-to-powder mass ratio increase in these films up to
-6.0 C/m
2
which is very close to our earlier work (-6.5 C/m
2
) of PZT thick films derived
by pure sol-gel process,20 but is somewhat less than the values (e.g. e
31,f
= -8.4 C/m
2
)
obtained by other workers in thick films produced by a pure sol gel process in 5 µm
PZT40/60 films.

48

Fig. 3.7 e
31,f
values of the PZT films with variable solution-to-powder mass ratios

To investigate the cause of above improvements in film properties, densities of
the films with different ratios were measured. Earlier, Nelson (2005) detailed various
sets of dielectric mixture equations considering the dielectric constant and density of
different materials and can often be used to predict the density of films from its dielectric
constant value or vice versa. In our case, an approximate linear relationship between film
densities and the square root of the film dielectric constants was noticed indicating a
complex refractive index mixture rule applies fairly well at various mass ratios. As
shown in the Fig. 3.8, the density of the fabricated film increased for a larger solution-to-
powder ratio in the composite solution. At a mass ratio of 0.5, the average density of the
film is only 5900 kg/m
3
, increasing to 6700 kg/m
3
at a mass ratio of 6 which is about
90% of bulk PZT-5H material (7500 kg/m
3
). The increase in e
31,f
values with higher
concentrations of loaded power mass ratios is definitely due to a decrease in film porosity

49
thereby increasing the d
31
(Bowen et al., 2004). Moreover, Maki et al. (2000) found that
the decrease in PZT thick film density leads to a reduction in the remanent polarization
values. Therefore, the improvements dielectric and piezoelectric properties of the PZT
film are mainly due to increased density with reduced porosity. The porosity of the film,
as discussed above, can be reduced by using a larger mass ratio composite solution
combined with a vacuum infiltration procedure.

**: mass ratio values
Fig. 3.8 Variation of film dielectric constants with film densities.

Our experimental results indicate that a mass ratio of around 4 is an optimal value
to obtain thick (>10 µm), dense and crack free PZT films with superior piezoelectric and
dielectric properties in this process. There are mainly three reasons: (i) The dielectric
constant (1200) and transverse piezoelectric values (–6 C/m
2
) in the films with mass
ratios around 4 are comparable to typical randomly oriented PZT thin film values (Maeda
et al., 1998; Corker et al., 2002). (ii) In the thick films with mass ratio greater than 4, the
(1)
(2)
(0.5) **
(4)
(6)

50
remanent polarization (Pr) values of the films were observed to be saturated. (iii) The
viscosity of the composite solution decreases with an increase in the proportion of PZT
sol gel. This results in that the higher the proportion of the PZT solution in the composite;
the larger the number of spin/infiltration steps would be needed to obtain thick films (~10
-20 µm). For example, each layer thickness was ~2 µm with a composite solution of
mass ratio equals to 1, whereas, only 0.5 µm thick film could be obtained with a mass
ratio of 4. This implies that the quality of the films will be more difficult to be maintained
with higher proportion of the PZT solution in the composite since each extra layer has the
risk of impairing the films quality. In addition, the higher dielectric constant of the thick
films can be obtained by loading high dielectric constant powder and using as same
composition sol-gel solution as ceramics.

3.2.4 Summary
In summary, it has been shown in this study that increasing of mass ratio of PZT
sol-gel solution to PZT powder in the composite solution gives rise to improved
piezoelectric and dielectric properties without using any liquid-phase sintering aids.
Particle Size Distribution (PSD) analysis experiments confirmed that particle size of the
composite solution has an insignificant role for the improvements. The improvements are
primarily due to homogeneity in the slurry composition that reduced both cracking and
elimination. Moreover, vacuum infiltration of sol-gel solution into films after each layer
deposition reduced void density. As a result, the overall film density increased with
enhanced piezoelectric and dielectric properties in thick PZT films. The improved

51
properties make PZT composite films promising candidates for high frequency
ultrasound transducers. Moreover, the relative low density of the PZT films (~70 % - 90
%) results in low acoustic impedances. This property makes the films better matching to
human tissues, and which is favorable to more suitable for thickness-mode broadband
transducers capable of medical imaging applications.
3.3 Characterization of PZT Thick Film
The finished PZT film was monitored using a scanning electron microscope
(SEM, Hitachi, S-3500N, Tokyo, Japan). Fig 3.9 (a) shows the particle size is about 200-
250 nm and thickness of the film as shown in Fig 3.9 (b) is around 12 µm with which is
suitable to build high-frequency (>100 MHz) transducers.

(a) (b)
Fig 3.9 SEM images of top view (a) and cross-section view (b) of a PZT film

The structure of the film was examined using a Rigaku X-ray diffractometer
(XRD). Fig 3.10 shows an XRD pattern of a PZT thick film deposited on Pt/silicon
substrate. The film appears to be well-crystallized and on strong peak along (110) was
observed.

52

Fig 3.10 XRD pattern of a PZT thick film

Several important parameters of the film were measured for modeling purposes.
To quantitatively characterize the properties of the film, circular Cr/Au electrodes with a
diameter of 1.5 mm were first deposited by sputtering.
Dielectric properties were measured using an Agilent 4294A impedance analyzer
and polarization - field (P-E) hysteresis properties were evaluated using a Radiant
precision materials analyzer. Fig. 3.11 and Fig. 3.12 show the ferroelectric hysteresis
loop and dielectric properties of the film, respectively. The film was found to have the
remanent polarization value of 33 µC/cm
2
, coercive field of 70 KV/cm, relative dielectric
constant of 1200 and loss of 0.03. The clamped dielectric constant was calculated to be
600.

53

Fig 3.11 Hysteresis loop of the PZT thick film

Fig 3.12 Dielectric properties of the a PZT thick film

The density of the film was also measured. In the study, the PZT film was first cut
into 10 mm × 10 mm squares by a dicing saw. The thickness of the film (over 10 µm)
was determined using a SEM (Hitachi, S-3500N, Tokyo, Japan) to allow the

54
determination the volume. Then, the parts were immersed in a KOH 20 % solution at a
temperature of 80 °C. This resulted in the PZT film peeling off from the silicon substrates
in approximately 10 minutes. The SEM images (Fig.3.13) show that there is no obvious
damage to the film when it was removed from the silicon substrate while causing slight
damage to the films (Frood et al., 2007). Finally, the mass of the film was measured using
a precision electronic balance (OHAUS Corp, Pine Brook, NJ) after clearing Pt/Ti/SiO
2

debris and drying. The density of the film was calculated to be 6300 kg/m
3
, which is
about 85 % of bulk PZT-5H material (7500 kg/m
3
).

(a) (b)
Fig 3.13 SEM pictures of a film surface after removing from the silicon by KOH

One of the most important parameters of a piezoelectric material is its
electromechanical coupling coefficient, which is a measure of the efficiency with which
the material converts mechanical energy to electrical energy or vice versa. Fig 3.14 (a)
shows the measured electrical impedance of a PZT films with a silicon substrate. The
oscillating in the plot is caused by the silicon. Lukacs et al. (1999) created a mathematical
model to predict coupling coefficients of a film clamped on a substrate. With the model,
the film’s thickness mode coupling coefficient (k
t
) was calculate to be 0.34. To validate

55
the result, the film was measured again after the silicon substrate was removed. This
time, the effective electromechanical coupling coefficient (k
t
) of the film was calculated
to be 0.31 according to IEEE standard (IEEE, 1987) with the measured electric
impedance curves shown in Fig 3.14 (b). The value is a little bit lower than the modeling
prediction, but it is quite reasonable considering that removal of the silicon will cause
crakes inside the film which in turn degrade the coupling factor.

(a)

56

(b)
Fig 3.14 Measured electric impedance curves of PZT thick films with silicon substrate (a)
and without silicon substrate (b).

The above measured properties were summarized and compared with PZT-5H in
Table.3.1. The data will be used in next chapter as pulse-echo modeling input for
PiezoCAD (Sonic Concepts, Woodinville, WA) software.
Table.3.1 Important properties of PZT film and PZT-5H
Material k
t
ε
33S
/ε
0
ρ (g/cm
3
) c (m/s) tan δ (%)
PZT film 0.34 600 6300 2900 3.4
PZT-5H 0.55 1470 7500 4560 2.0
k
t
is the thickness mode electromechanical coupling coefficient; ε
33S
/ε
0
is the clamped
dielectric constant; ρ is the density, c is the longitudinal wave velocity and tanδ (%) is the
dielectric loss.

57
3.4 Poling of PZT Thick Film
The PZT bulk materials and PZT films differ in two major properties: films
exhibit much higher coercive fields (typically 5 V/µm to 10 V/µm) and higher
breakdown voltages (20 to 40 V/µm). Therefore, it is necessary to drive films with
higher fields in order to compensate partially for the smaller thickness. Depolarization
takes place when the operation field is too large compared to the coercive field.
To acquire the pulse-echo signal, the PZT thick film had to be properly polarized
before assembling the array. Wang et al., (1989) and Takahiro et at., (2000) reported and
discussed the poling conditions of PZT ceramics. But those poling conditions are not
suitable for the PZT films. The poling electric field of PZT films should be much higher
than that of PZT ceramics (normally 2~ 3 volts/µm) (Damjanovic, 1998). A set of
experiments were performed to determine the optimal poling conditions of the PZT
composite films. In the poling process, the poling electrical field, temperature and
duration are variable. First, the oven was set at a fixed temperature. Variations of
electrical fields from 2 V/µm to 16 V/µm and poling durations of 5 minutes to 30 minutes
were applied to find an optimal poling condition. Subsequently pulse-echo signals were
recorded from the samples after they were cooled down to room temperature. We stop the
experiment until no obvious improvement was observed with a higher temperature set.
The experiment suggested the poling duration of 10 minutes is longer enough for the PZT
films, the measured parameter is saturated with longer poling duration. Furthermore, as
shown in Fig. 3.15, under constant temperatures, the magnitude of the output voltages of
the poled PZT film transducer increased with increasing poling fields and reached a

58
nearly constant value of at around 12 V/µm; when the poling fields were kept constant,
the optimal poling temperature to reach maximum output voltage was around 160 ºC.

Fig 3.15 Electric field effects and temperature effects in poling process

59

CHAPTER 4
FABRICATION OF PZT THICK FILM SINGLE-ELEMENT TRANSDUCERS
AND KERFLESS ARRAYS

Fabrication of high-frequency (between 30 MHz and 50 MHz) ultrasonic linear
arrays is still a challenge. The task is even more difficult to build arrays at a frequency
greater than 100 MHz, which has the potential to provide more detail skin texture for
early diagnosis of melanoma or to image small objects such as stem cells. Recent
research shows that MEMS micro-machined technology is a possible solution to such
high-frequency applications although much improvement still needs to be made
(Bernstein et al., 1997; Pang et al., 2006; Mina et al., 2007; Dauchy et al., 2007).
Piezoelectric films have already been widely used to fabricate micro-scale devices
(Muralt, 2000). In previous chapter, PZT film has been shown to exhibit excellent good
dielectric and piezoelectric properties, making it a good possible candidate for ultrasonic
transducer material (Xu et al., 2002; Trolier-McKinstry et al., 2004). In recent years, PZT
film has been used to fabricate high-frequency single-element transducers, kerfless
annular arrays, lateral mode linear arrays and low-frequency 2-D arrays (Lukacs et al.,
2000; Robert et al., 2007; Ioanna et al., 2007; Dausch et al., 2006). Although PZT films
typically have a lower electromechanical coupling coefficient than bulk PZT, it has they
have several advantages over conventional PZT bulk materials in ultrasonic transducer
applications. For example, the PZT film has relatively low acoustic impedances (15~20
MRayls) compared to conventional PZT bulk material (~ 30 MRayls), which is useful in

60
the fabrication of broad bandwidth ultrasonic transducers. In addition, if the sintering
temperature of the PZT film can be further decreased in the future, the PZT film
transducers and their processing circuits could be fabricated at on the same silicon chip.
A thickness mode kerfless high-frequency PZT film linear array was developed as an
alternative to kerfed arrays if a higher cross-talk level is acceptable, since dicing fine
kerfs at present is still a very expensive endeavor. This issue has been addressed by
Morton et al. (2002) and Demore et al. (2002). Performance of the kerfless array
prepared from PZT films was compared to a PZT-5H ceramic sheet array fabricated from
an identical array pattern.

4.1 Fabrication of PZT Single-element Transducers
4.1.1 Modeling
PiezoCAD (Sonic Concepts Inc., Woodinville, WA), based on Krimholtz,
Leedom and Matthaei (KLM) model (Krimholtz et al., 1970) was used help designing the
transducer. The model included a 12-µm PZT film, a 0.3 mm diameter aperture. E-solder
3022 was used as backing. No matching layer was added. A pulse echo response from
the model shown in Fig.4.1, displays the 17.8 nano-second, -6-dB pulse length. The
transform of the pulse echo response simulation shows a center frequency of 119 MHz
with 38 % bandwidth. At 119 MHz, the modeled impedance magnitude and phase were
40 Ω and 35° respectively.

61
0 0.07 0.14 0.21 0.28
-15
-10
-5
0
5
10
15
Time (us)
Vo/ Vi (mV/V)

60 90 120 150 180
-30
-24
-18
-12
-6
0
6
Frequency (MHz)
Vo/ Vi (dB)
Magniture
Spectrum

Fig 4.1 Modeling pulse-echo response of the PZT film single element transducer

4.1.2 Fabrication Process

Fig 4.2 Structure of the PZT film single element transducer

62
First, a single element transducer was built with the 12-µm PZT thick film.
Circular electrodes were sputtered on the surface of the thick films with a home-made
pattern. The area of the electrodes was sized for 50-Ω electrical impedance. The silicon
substrate was etched out with KOH solution at 80 °C. The film was then glue to E-solder
3022 with non-conductive epoxy. SMA connector was finally mounted. The structure of
the single elopement transducer is shown in Fig 4.2.

4.2 Fabrication of PZT Kerfless Arrays
A mask was designed for patterning the 32-element linear arrays. In our initial
design, the linear array had a kerf of 12 µm, an element width of 24 µm and an element
length of 4 mm as shown in Fig.4.3.

Fig 4.3 Structure of the PZT film single element transducer

63
4.2.1 Modeling
With data list in Table 3.1, PiezoCAD software was used to predict the
performance of the kerfless arrays. In the models, E-solder 3022 (5.9 MRayls) was
selected as the backing material. No matching was added. The modeled results predicted
that the center frequencies of the PZT film and PZT-5H array are 118 MHz and 132 MHz,
as shown in Fig.4.4 (a) and (b), respectively. The modeled results also suggested that
without any front matching, the bandwidth of the PZT film array may reach 47 %, which
is 17 % higher than PZT-5H array (30 %).

0 0.025 0.05 0.075 0.1
-3
-1.5
0
1.5
3
Time (us)
Vo/Vi (mV/V)

70 95 120 145 170
-24
-18
-12
-6
0
Frequency (MHz)
Vo/Vi (dB)
Voltage
Spectrum

(a)

64
0 0.025 0.05 0.075 0.1
-3
-1.5
0
1.5
3
Time (us)
Vo/Vi (mV/V)

70 95 120 145 170
-24
-18
-12
-6
0
Frequency (MHz)
Vo/Vi (dB)
Voltage
Spectrum

(b)
Fig 4.4 Modeling pulse-echo response of PZT film (a) and PZT-5H ceramic (b)
kerfless array elements

4.2.2 Fabrication Process
The details of the fabrication processing are illustrated in Fig. 4.5. First, an
insulating layer of 0.1-µm Si
3
N
4
was deposited onto PZT thick film and patterned to
leave an opening for the working area of the linear arrays. Then a layer of Au/Cr was
deposited on the top surface and patterned as the top electrodes with photolithography. If
the Au/Cr electrode is too thick, it will damp the center frequency. Connection will be a
problem if the Au/Cr layer is too thin. Based on our experience, 1000Å/500Å of Au/Cr is
a reasonable thickness: it can provide good conduction but only damp ~ 1 MHz of the
center frequency. The backside silicon under the working area was etched with a XeF
2

65
etching system. Subsequently, the layer of SiO
2
was removed by BOE. Finally a
conductive epoxy, E-solder 3022 (Von Roll Isola Inc., New. Haven, CT), was used as
filler and provided the electric connection for the backside electrode, and at the same
time, acted as a backing material. Fig. 4.6 shows the pictures of the device fabricated.
Only in the working area (marked in green), are the top electrodes directly in contact with
PZT thick film. All other areas are covered by the insulating layer of Si
3
N
4
.
The fabrication of PZT-5H sheet array is very similar to the PZT thick film array.
The only major difference is that PZT-5H sheet was used as the piezo-material in this
experiment because the powder used to fabricate the PZT thick film is PZT-5H. A piece
of bulk PZT-5H sheet was bonded onto a silicon substrate with E-solder 3022 and was
later lapped to 15 µm thickness with a precision lapping machine (DAG 810 Grinder,
DICSO). The same mask was used to pattern the kerfless arrays with the same micro-
fabrication procedures described previously.

66

Fig 4.5 Fabrication process for PZT film kerfless arrays

Fig 4.6 Pictures of top electrodes of the thick film linear arrays with insulation layer
underneath.

67
4.3 Results and Discussion
It is difficult to measure the very high frequency PZT thick film transducers.
Selection of pulse and receiver is critical for acquiring pulse-echo signals from the PZT
thick-film transducers. Several popular pulser and receiver instruments were evaluated.
Panametrics 5900PR combines a pulser and a receiver together, providing the
working frequency up to 200 MHz. Fig.4.7 show measured pulser signals and their
spectrums of the 5900PR with energy setting of 1, 2, 4 µJ respectively. As shown in Fig
4.7 (b) and (c), with energy setting of 2, 4 µJ, most energy of the pulser signals is located
at very low frequency. The pulser’s performance deteriorates severely with high energy
settings. It is not suitable to measure pulse/echo with high energy settings for high-
frequency (> 50 MHz) PZT thick film transducers anymore.
-0.053 -0.028 -0.003 0.022
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

0 37.5 75 112.5 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(a)

68
-0.053 -0.028 -0.003 0.022
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

0 37.5 75 112.5 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(b)

-0.053 -0.028 -0.003 0.022
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

0 37.5 75 112.5 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(c)

Fig 4.7 Pulse plot of 5900PR with 1 µJ (a), 2 µJ (b) and 4 µJ (c) setting

69
Panametrics 5910R combined with 5910RPP is a good pulser candidate,
especially with energy setting at LOW. Fig.4.8 show 5910’s pulses and their spectrums
with energy setting at LOW and HIGH respectively. But the receiver part of the 5910 is
very noisy. Broadband (5-400 MHz), low noise figure (1.2 dB) Miteq AU1466
Preamplifier with Gain of 35 dB thus was selected as the receiving unit. Besides
Panametrics instruments, Avtech AVB2-TC-C monocycle generator can also be used as
the pulser for our high-frequency (>100 MHz) transducers. Fig 4.9 shows its pulse and
spectrum with setting of 100 MHz in frequency and 130 Volt in amplitude. The
advantage of the Avtech is their pulses’ center frequency and amplitude is adjustable.
-0.027 -0.0145 -0.002 0.0105
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

0 37.5 75 112.5 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(a)

70
-0.027 -0.0145 -0.002 0.0105
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

0 37.5 75 112.5 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(b)
Fig 4.8 Pulser plot of 5910 with LOW (a) and HIGH (b) energy setting

-0.049 -0.024 0.001 0.026
-3
-1.5
0
1.5
3
Time (us)
Voltage (V)

50 75 100 125 150
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

Fig 4.9 Avtech AVB2-TC-C monocycle generator’s pulse and spectrum
(With center frequency of 100 MHz and amplitude of 130 volt)

71
4.3.1 Single-element Transducer
The pulse echo is measured using a water bath and quartz target. In this
measurement Avtech AVB2-TC-C monocycle generator was selected as a pulser. Miteq
AU1466 Preamplifier was used as a receiver unit. Data is visualized on a LeCroy LC534
digital oscilloscope (Fig 4.10).
Reasonable pulse-echo signals were successfully acquired from PZT film single
element transducers and kerfless arrays. Fig. 4.11(a) shows the pulse-echo plot of a
single element transducer. The single element transducer shows the center frequency of
120 MHz, -6-dB bandwidth of 40 % and peak-to-peak voltage of one volt with 35 dB
gain.

Fig 4.10 Pulse-echo test set-up

Oscilloscope
Avetech pulser/ Miteq preAmp
PZT film transdcuer
Quartz Target
Water Bath
50 Ω Cables

72
0.706 0.756 0.806 0.856 0.906
-1
-0.5
0
0.5
1
Time (us)
Voltage (V)

70 95 120 145 170
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(a)

1 1.05 1.1 1.15 1.2
-1
-0.5
0
0.5
1
Time (us)
Voltage (mV)

50 80 110 140 170
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(b)

Fig 4.11 Measured pulse-echo plots of a PZT film single-element transducer before (a)
and after (b) deposting parylene as a matching layer

73
To improve the performance of the transducer and array, 7-µm parylene were
deposited as the matching layer. -6-dB bandwidths of the single-element transducer
increase to 60 % after the parylene deposition as shown in Fig 4.11(b).

4.3.2 Kerfless Arrays
After proper poling and packaging, the pulse-echo signals of the PZT film and
PZT-5H arrays were successfully recorded and plotted in Fig.4.12 (a) and (b),
respectively. The measured results pulse-echo of PZT film array is in good agreement
with the modeled data. This confirms that the measured PZT film properties in Table3.1
are accurate. The PZT-5H array’s echo shows less rings than the modeling result. Most
likely, it was caused by either the deviation of data values provided in the PZT-5H
datasheet or the fabrication errors such as the acoustic impedance of the E-solder is
higher than the experimental value of the 5.9 MRyals. To further improve the bandwidth
of the arrays, a layer of 5-µm parylene was deposited onto the front-surface of the arrays
as the matching material. With the parylene matching, the bandwidth of the PZT film
array might be as high as 60% and the bandwidth of the PZT-5H array increased to 40%.
Higher bandwidth of the PZT film array is attributed to its lower acoustic impedance
compared to PZT-5H sheet array.

74

(a)

2.12 2.145 2.17 2.195 2.22
-350
-175
0
175
350
Time (us)
Voltage (mV)

70 95 120 145 170
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(b)
Fig 4.12 Measured pulse-echo response of PZT film (a) and PZT-5H ceramic (b)
kerfless array elements

75

In order to compare the sensitivity of the arrays, two-way insertion loss and cross-
talk between nearest adjacent elements measurements were performed. In the setup, a
piece of 2-inch quartz was immersed in a degas water bath at room temperature as a
target. 20-cycle 5-V sinusoid high-frequency bursts was generated by Tektronix
AFG2020 function generator (Tektronix, Inc, Richardson, TX) as the signal sources. The
echo signals were acquired and displayed using the LeCroy LC534 1GHz digital
oscilloscope. Because the echo signal of the thick film array is weak, the echo was first
amplified by a 35-dB low-noise high-frequency Miteq AU-1466 (Miteq, Hauppauge, NY)
preamplifier before feeding to the oscilloscope. Losses due to attenuation in the water
bath and the transmission of the signal in the quartz target were compensated but the
losses caused by diffraction were not included. Fig.4.13 shows the pulse-echo, insertion
loss and cross-talk plots of a PZT film element after depositing parylene as matching. All
of the above measured results are summarized in Table.2. The PZT film array and PZT-
5H array was found to have insertion losses of 41 dB and 28 dB, respectively. The
results showed that the PZT film kerfless array had a center frequency of 120 MHz, a
bandwidth of 60% with a parylene matching layer and insertion loss of 41 dB. The PZT
sheet kerfless array was found to have a center frequency of 128 MHz. The measured
results revealed that the PZT sheet array has a poor bandwidth (40 % with a parylene
matching layer) but a better sensitivity (28 dB insertion loss).
The measured results also suggest that both the kerfless arrays have cross-talks in
between -20 dB and -15 dB, which are relatively high. In the following chapter, DRIE

76
(Deep Reactive Ion Etching) dry etching technology will be used to separate the arrays’
elements with. This will decrease the cross-talks between elements.

Table.4.1 Modeling and measured results of representative PZT thick film and PZT-5H
ceramic elements
Material Center Frequency
(MHz)
Bandwidth (%)
no matching
Bandwidth (%)
with matching
Insertion
loss (dB)
PZT film 120 (120)* 45 (47) 60 (62) 41 (33)
PZT-5H 128 (123) 32 (30) 40 (44) 28 (21)

*: the values insider the parentheses are the Piezo-CAD modeling results.

1 1.05 1.1 1.15 1.2
-300
-150
0
150
300
Time (us)
Voltage (mV)

50 80 110 140 170
-24
-18
-12
-6
0
Frequency (MHz)
Spectrum (dB)
Voltage
Spectrum

(a)

77

(b)

(c)

Fig 4.13 Pulse-echo (a), insertion loss (b) and cross-talk plots of a PZT film kerfless
element with parylene as a matching

78
4.3.3 Ultrasonic Imaging with Film Single-element Transducer
20-µm diameter Tungsten wires were first used to assess the spatial resolution of
the PZT single element transducer. Fig 4.15 and Fig 4.16 shows the wire phantom image
with 12 dB and 6 dB dynamic ranges respectively. The single element transducer was
found to have an axial resolution (-6 dB) of 20 µm and lateral resolution (-6 dB) of 100
µm.

(a)

79

(b)

Fig 4.14 Image of wire targets with 12 dB (a) and 6-dB dynamic range

Ultrasonic image of a porcine eyeball was also obtained with the single element
transducer as shown in Fig.4.17. In the image, the cornea is clearly shown, but the lens is
barely seen at this high frequency. Fig 4.18 shows ultrasonic image of a normal human
skin. In the image, follicle can be distinguished from surrounding tissues.

80

Fig 4.15 Ultrasonic image of a normal porcine eyeball

Fig 4.16 Ultrsonic image of normal human skin

Cornea
Lens
Follicle

81
CHAPTER 5
FABRICATION OF PZT THICK FILM KERFED ARRAYS
In last chapter, we demonstrated the feasibility of fabrication very high frequency
kerfless linear array. However, the measured results show cross-talks between adjacent
elements are very high (21 dB). To reduce the cross-talks, mechanical isolation between
elements is necessary. Conventionally, arrays are manufactured by a “dice-and-fill”
approach, where a piezoelectric plate is separated by mechanical dicing and a polymer is
infiltrated and cured within the kerfs. This method, however, cannot be used to make
transducer arrays at frequencies higher than 50 MHz. In this chapter, we will investigate
a practical approach to producing kerfed linear arrays by MEMS dry-etching technique.

5.1 Dry-etching Technique
Deep reactive ion etching (DRIE) has been widely used in etching PZT thin films
for ferroelectric memory applications (Kokaze, et al. 2007). Recently, new advances in
PZT dry etching have been reported (Bale et al., 2001; Wang et al., 1999; Subasinghe et
al., 2006). They use pure sulfur hexafluoride (SF
6
) or a mixture of SF
6
and Argon (Ar) as
etching gases in the plasma etching of PZT. The advantages of fluorine chemistry are its
good selectivity to mask materials and its relatively high etch rate, which can reach
values as high as 0.25 µm/min. However, because the components of the PZT etch
products have significantly higher boiling temperatures their removal is inefficient. In
closely spaced structures, this results in sidewall angles < 80°, which is unacceptable for
fabricating high frequency arrays. On the other hand, Marks et al. (2003) reported that

82
the use of chlorine-based etching gases with an elevated wafer temperature yields good
vertical etch profiles and a higher etch rate. Also the dry etching mechanism of PZT in
chlorinated gas plasmas has been investigated in detail (Jung et al., 2001; Efremov et al.,
2004), and some measures have been proposed to minimize etching damage to the PZT
material during the dry etch process (Kang et al., 2002). In this research, process
parameters for the chlorinated gas plasma have been optimized and applied to PZT film
etching.

5.2 Modeling
The mask used to pattern the kerfed arrays is the same as the kerfless one: it has
element number of 32, the kerf of 12 µm, and the element width of 24 µm and element
length of 4 mm. To validate the feasibility of the array, both PZFLEX and Field II
modeling were carried out. The PZFLEX program was used to predict the azimuthal
beam profile for eight elements in order to investigate the impact of grating lobes. A
transmitting grating lobe was expected near 40º as shown in Fig.5.1 (a). The feasibility
of imaging with the array was simulated using a Field II program. A phantom with five
point scatter was chosen as the target. The result in Fig. 5.1 (b) shows that, with focusing
delay, the array can be well focused at 1-2 mm.

83

(a) (b)

Fig 5.1 FIELD II (a) and PZFLEX (b) modeling of the linear array

5.3 Fabrication Process
A. Hard mask deposition
A photo-lithography-based method was applied to transfer design patterns onto
the film surface. A layer of Cr/Au (500 Å/1000 Å) was first patterned onto the
film surface as the seed layer. Then a thick positive photoresist layer was spin-
coated on top of the seed layer. After baking on a hot plate, the photoresist was
exposed directly by a laser writer. The wafer was dipped in photoresist developer
for developing. The patterned photoresist structures were then formed. Next, the
Ni electroplating process was run to form a 4-µm thick Ni hard mask through the
openings of the photoresist pattern. In order to get the dense Ni layer in the small
open areas, a pulse power supply (Dynatronix, Amery, WI) was used to provide

84
pulses with 0.1 ms on and 0.9 ms off, and with the average current 20 mA/cm
2
.
After electroplating, the photoresist was stripped off using Acetone.
B. Dry etching
ICP (Inductively Coupled Plasma)-RIE (Reactive Ion Etching) dry etching
technique was used to etch the film. The area underneath the nickel hard mask
was protected; the rest area was etched off by chlorine gas plasma. Etching rate,
selectivity ratio, and profile angle are the most important parameters when dry
etching PZT film for the construction of high-frequency ultrasound transducer
arrays. By optimizing ICP power, substrate power (controlling substrate DC bias
voltage), flow rates and ratios of etching gases, we found that a median etching
rate of 8 µm /hr was a suitable value for the process. It takes too much time to
achieve the final thickness at lower etch rates, and higher etch rates lead to
deterioration of the profile angle. Finally we used an inductively coupled plasma
reactive ion etching system (Plasmatherm SLR770 ICP system, Unaxis Inc.,
Irvine, CA) employing Cl
2
/Ar based chemistry for the etching process. By using
600 W of ICP power, the DC bias voltage of 200 V, 20 sccm of Ar, and 60 sccm
Cl
2
, the etching rate of PZT film was 8 µm /hr. Fig. 5.2 shows the SEM photos of
the etched PZT film. The measurements show that the thickness of the elements is
~18 µm, the sidewall angle of the elements >85°.

85

Fig 5.2 SEM picture of a linear array dry-etched from a PZT film

C. Filling kerfs
The kerfs were infiltrated with Epo-Tek 301-2 epoxy (Epoxy Technology,
Billerica, MA). Before filling, MARCH plasma cleaning system was conducted to
make the array surface hydrophilic, so that the epoxy is easier to be filled into the
micro-scale kerfs. Degassing process was conducted right after filling. The epoxy
was cured at the room temperature over night followed by 2 hours curing at an
elevated temperature of 60 °C. After curing, the extra epoxy and residual nickel
were carefully lapped away with fine sandpapers followed by 3-µm Al
2
O
3
powder.
D. Patterning electrodes
A layer of Si
3
N
4
(1000 Å) was deposited as the insulating materail covering the
electrodes area of the array before a layer of Cr/Au (500 Å/1500 Å) electrode was
patterned. Fig. 5 (a) and (b) shows the picture of a PZT film kerfed linear array
before and after epoxy filling, respectively. The dark area in Fig. 5 (b) is the
place where the Si
3
N
4
was deposited.

86

(a) (b)
Fig 5.3 Photography of the array before and after epoxy filling

E. Substrate removal
The backside silicon under the working area was etched with a XeF
2
etching
system and the SiO
2
was removed by BOE. Subsequently, conductive epoxy, E-
solder3022, was filled, providing the electric connection for the backside
electrode, and at the same time, acting as a backing material.
F. Interconnection
A flexible circuit (Fig. 5.4) and a PCB connector board (Fig.5.5) were designed to
connect array elements electrically with high reliability and low noise level. Each
flexible circuit connects 16 array elements. The narrow side of the flexible circuit
was aligned and bonded to array elements; the wide side was aligned and bonded
to the connecting pads of the PCB board. Micro-coaxial cables were soldered onto
the small holes in the PCB board. In the PCB connector board, the central area
and blue lines, as shown in Fig. 5.5 (a) are ground.

87

(a) (b)

Fig 5.4 Layout (a) and photography (b) of the flexible circuit

(a) (b)

Fig 5.5 Layout (a) and photography (b) of the PCB connector board

G. Poling
The poling process was carried out after patterning the electrodes and before
interconnection. There are couples of reasons for that. First, we would like to
avoid any heat treatments performed after poling. Those heat treatments will

88
do-poling the PZT films. Electrode pattering will be involved in 90°C heat
treatments to bake the photo-resist. There are no more heat treatments after that.
But, we cannot perform the poling process after interconnection and packaging.
The bonding epoxy, which was used to bonder the flexible circuits, will be
damaged at such high poling temperatures.
A kerfed PZT thick-film array was successfully fabricated with the above
procedures. But there are still several problems with the fabrication method. Firstly, the
array is very likely to crack when the silicon substrate was removed and E-solder was
filled. Secondly, with the method, the area underneath the electrodes has to be coved by
insulating material. Therefore, a majority part of the film was wasted. Lastly, the array is
not possible to be press-focused because of the hard silicon substrate underneath the film.
Thus a modified fabrication process is suggested.
In the new method, E-solder 3022 was first casted onto the front surface of the
PZT thick film and centrifuged 5 minutes at 3000 RPM. After curing the E-solder over
night, the PZT film was cut into 4 mm × 2 mm parts by a dicing saw. Then, the parts
were immersed in a KOH 20% solution at a temperature of 80 °C. This resulted in the
PZT film peeling off from the silicon substrates and still firmly bonding to the E-solder.
Moreover, we have demonstrated in last chapter that there is no obvious damage to the
film when it was removed from the silicon substrate. Place a plastic tube concentrically
around the sample. Epotek-301 non-conductive epoxy was filled into the tube and
degassed to remove the air bubbles. The film surface was sealed by Kapton tap, avoiding
touching the epoxy. After curing, the extra epoxy was lapped away. Fig. 5.6 shows the

89
top view and bottom view of the array sample before interconnection. The following
procedures are same to the previous method except that the insulating layer patterning
and silicon etching are not necessary. The finished array is shown in Fig.5.7.

(a) (b)
Fig 5.6 Top view (a) and bottom view (b) of the kerfed array before innterconneciton

Fig 5.7 Picture of the array after innterconncetion

90
5.4 Results and Discussion
The measurement set-ups are the same as the kerfless arrays’, which was
described in last Chapter..
The measured pulse-echo plot, as shown in Fig. 5.8, shows that, after depositing a
layer of parylene as a matching, the array has the center frequency of 80 MHz and -6 dB
bandwidth of 80%. The broad bandwidth mainly benefit from the low acoustic
impedance of the PZT film.

Fig 5.8 Measured pulse-echo of the kerfed array

Fig.5.9 shows the uniformity of the array elements’ capacitance and dielectric loss.
As shown, over all, the array is quite uniform: the capacitance is located around 30 pF;
the dielectric loss is around 0.04. Most importantly, there are no broken elements

91
observed. Fig.5.10 shows measured insertion loss of a reprehensive array element. After
compensation for the attenuation caused by the water and reflection from the quartz
target, an insertion loss of -41 dB was obtained at the frequency of 90 MHz. The
relatively high insertion loss is mainly due to small coupling coefficients of the PZT films.
There is still a large room to further improve PZT film’s functional properties making it
comparable to PZT bulk materials.
The results of combined acoustic/electrical cross-talk measurement are presented
in Fig 5.11. A set of results for nearest neighbor are measured. A crosstalk of around -21
dB was observed between 50 MHz and 90 MHz, which is around 5 dB lower than the
PZT film kerfless array.

Fig 5.9 Measured element uniformity of the kerfed array

92

Fig 5.10 Measured insertion-loss of a typical kerfed array element

Fig 5.11 Measured cross-talk between adjacent kerfed array elements

93
CHAPTER 6
SUMMARY AND FUTURE WORK
6.1 Summary of Results
This study shows that it is a promising approach to fabricate a high frequency
(~100 MHz) single element transducers and linear arrays using PZT thick films. This
simple yet effective approach for single-element transducers and arrays fabrication
proved to be reliable and reproducible, and is well suited for mass production of future
devices at high frequency.
The performances of the PZT film ultrasonic transducers are ultimately determined
by the functional properties of the PZT films. To improve the properties of the PZT films,
optimized fabrication procedures were suggested. It has been shown in the study that
increasing of mass ratio of PZT sol-gel solution to PZT powder in the composite solution
gives rise to improved piezoelectric and dielectric properties without using any liquid-
phase sintering aids. Particle Size Distribution (PSD) analysis experiments confirmed that
particle size of the composite solution has an insignificant role for the improvements. The
improvements are primarily due to homogeneity in the slurry composition that reduced
both cracking and elimination. Moreover, vacuum infiltration of sol-gel solution into
films after each layer deposition reduced void density. As a result, the overall film density
increased with enhanced piezoelectric and dielectric properties in thick PZT films. The
improved properties make PZT composite films promising candidates for high frequency
ultrasound transducers. Moreover, the relative low density of the PZT films (~70%-90%)
results in low acoustic impedances. This property makes the films better matching to

94
human tissues, and which is favorable to more suitable for thickness-mode broadband
transducers capable of medical imaging applications.
Poling condition of the PZT films is different that of the PZT ceramics. Therefore,
a set of experiments were performed to determine the optimal poling conditions of the
PZT composite films. The results revealed that the poling electric field of PZT films (10-
16 V/µm) is much higher than that of PZT ceramics (normally 2~ 3 volts/µm).
The single-element PZT film transducer was built and evaluated. The measured
results showed that the single-element transducer has a bandwidth of 40 % without any
matching. The transducers bandwidth was improved by depositing a layer of parylene as
matching. With a layer of matching, bandwidth of the single-element transducer
increased to ~ 60 %. The high frequency and broad bandwidth features of the transducer
makes it well suited for high frequency medical imaging applications. A wire phantom
image was acquired to evaluate the spatial resolutions of the transducer. The transducer
was found to have an axial resolution of 20 µm and a lateral resolution of 100 µm.
Porcine eyeball and human skin imaging was also successfully obtained with the
transducer.
Thirty-two-element kerfless ultrasonic linear arrays were fabricated with PZT film
and PZT-5H sheet by lithography method. The experimental results suggest that PZT film
arrays may be useful in broad bandwidth applications because of its relatively low
acoustic impedances (15-20 MRayls) compared to PZT sheet materials (~ 30 MRayls).
PZT 5H sheet, on the other hand, has greater sensitivity. Because of much higher coupling
coefficient of the PZT-5H (k
t
= 0.55) than PZT film’s (k
t
= 0.34), the insertion loss of the
PZT-5H array is 13 dB lower than PZT film array’s.

95
Kerfed array was achieved using DRIE dry-etching method. In this research,
process parameters for the chlorinated gas plasma have been optimized and applied to
PZT film etching. The resulted array shows the element thickness of ~18 µm, the
sidewall angle of > 85°. Conventionally, only the substrate underneath the array element
was etched. There are many problems with the method: the array is very likely to crack
when the silicon substrate was removed and E-solder was filled. Secondly, with the
method, the area underneath the electrodes has to be coved by insulating material.
Therefore, a majority part of the film was wasted. Lastly, the array is not possible to be
press-focused because of the hard silicon substrate underneath the film. Thus a modified
fabrication process is suggested. With the new method, the PZT was firstly transferred
from the silicon substrate to E-Solder material and then be cut into small pieces. The
finished arrays were evaluated. The results show that the cross-talk PZT film kerfed array
is 5 dB lower than the kerfless one. Their other performances, such as bandwidth and
insertion loss are similar.

6.2 Suggestions for Future Work
Many improvements may be made to the existing design and fabrication of the
PZT thick film transducers described in this work.
There is still room for further improvement of PZT film functional properties
making them comparable to PZT sheet by optimization of spin-coating procedure (such as
spin-coating speeds and durations), sintering condition, PZT pure solution vacuum filling
procedure (such as vacuum pressure, duration and filling sequence) , and PZT pure

96
solution to PZT powder mass ratio. Doping of PZT solution also has been shown to
enhance properties of the films.
To improve the lateral resolution of the PZT film single element transducer and
arrays, press-focus method is a practical approach to this problem. Especially for the linear
arrays, since silicon substrate can be safely removed, press-focused method now becomes
possible. Current version of the linear array only has 32 elements; the image aperture is
quite narrow. 128 or even 256-element linear arrays will produce images with much
improved qualities. Bonding the flexible circuits with epoxy is not a very reliable
approach. Moreover, the flexible circuit is a source of electric cross-talk. Ultrasonic wire
bonding has been widely used to connector miniature devices. It can be used in the future
to connect the array elements to PCB board directly with micro-wires.
There is still no ultrasonic imaging system available which can work at as high as
100 MHz for our arrays. It is suggested that synthetic aperture imaging be used to acquire
ultrasonic images with the linear arrays in the near future.

97
BIBLIOGRAPHY
Akedo J, Lebedev M, “Aerosol Deposition of Ceramic Thick Films at Room
Temperature: Densification Mechanism of Ceramic Layers,” Appl. Phys. Lett, 77 [11],
pp. 1710-1712 (2000).

ANSI/IEEE Standard on piezoelectricity, Std. pp.176-1987, IEEE, New York, (1987).

Bale M and Palmer R, “Deep plasma etching of piezoelectric PZT with SF6,” J. Vac. Sci.
Technol., B 19(6), pp.2020-2025 (2001).

Barrow DA, Petroff TE, Sayer M, “Thick ceramic coatings using a sol gel based ceramic-
ceramic 0-3 composite,” Surface and Coatings Technology., 76-77, pp. 113-118 (1995).

Barrow DA, Petroff TE, Tandon RP and Sayer M, “Characterization of thick lead
zirconate titanate films fabricated using a new sol gel based process, Journal of Applied
Physics, 81(2), pp. 876-881 (1997).

Bernstein JJ, Finberg SL, Houston K, Niles LC, Chen HD, Cross LE, Li KK,
Udayakumar K, “Micromachined high frequency ferroelectric sonar transducers,” IEEE
Trans. Ultrason. Ferroelect. Freq. Contr., 44, pp. 960-969 (1997).

Bharadwaja SSN, Olszta M, Dickey EC, and Trolier-McKinstry S, Li X, and Mayer T,
and Roozeboom F, “Fabrication of High Aspect Ratio Ferroelectric Microtube Structures
using a Vacuum Infiltration Technique,” J. Am. Ceram. Soc., 89 (9) pp. 2695-2701
(2006).

Biggers JV, Shrout TR, Schulze WA, “Densification of PZT Cast Tape by Pressing,”
Ceram. Bull., 58 [5], pp. 516-521 (1979).

Bowen CR, Perry A, Lewis A, Kara H, “Processing and properties of porous
piezoelectric materials with high hydrostatic figures of merit,” J. Euro. Ceram. Soc., 24,
pp. 541-545, (2004).

Cannata JM , Ritter TA, Chen WH, Silverman RH, and Shung KK, “Design of efficient,
broadband single-element (20-80 MHz) ultrasonic transducers for medical imaging
applications,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 50, pp.1548-1557
(2003).

Cannata JM, Williams JA, Zhou QF, Ritter TA, Shung KK, “ Development of a 35-MHz
piezo-composite ultrasound array for medical imaging,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr., vol. 53, pp.224-236 (2006).

98
Chang JF, Desu SB, “Effects of dopants in PZT films,” Journal of Materials Research.,
vol. 9, pp. 955-969, (1994).

Corker DL, Zhang Q, Whatmore RW, Perrin C, “ PZT ‘composite’ ferroelectric thick
films,” Journal of the European Ceramic Society., 22, pp. 383-390 (2002)

Damjanovic D, “Ferroelectric, dielectric and piezoelectric properties of ferroelectric thin
films and ceramics”, Rep Prog Phys, vol 61, pp. 1267-1324 (1998).

Dauchy F, Dorey FA, “Thickness mode high frequency MEMS piezoelectric micro
ultrasound transducers,” J. Electroceram., vol. 19, pp. 383-386 (2007).

Dausch DE, Castellucci JB, Chou DR and von Ramm OT, “Piezoelectric micromachined
ultrasound transducer (pMUT) arrays for 3D imaging probes,” IEEE Ultrasonics
Symposium., pp. 934-937 (2006).

Demore CM, Brown J, Lockwood GR, “Investigation of cross talk in kerfless annular
arrays for high-frequency imaging,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr.,
vol. 53, pp.1046-1056 (2006).

DeSilets CS, Fraser JD, and Kino GS, “Design of efficient broadband piezoelectric
transducers,” IEEE Trans. on Sonics and Ultrasonics, vol. 25, pp. 115-125 (1978).

Dey SK, Budd KD and Payne DA, “Thin-Film Ferroelectrics of PZT by Sol-Gel
Processing,” IEEE Trans. Ultrason. Ferroelect. Freq. Contr., 35, pp. 80-81 (1988).

Dorey RA, Stringfellow SB, Whatmore RW, “ Effect of sintering aid and repeated sol
infiltrations on the dielectric and piezoelectric properties of a PZT composite thick film,”
Journal of the European Ceramic Society., 22, pp. 2921-2926 (2002).

Duval FC, Dorey RA, Wright RW, Huang Z, Whatmore RW, “Fabrication and modeling
of high-frequency PZT composite thick film membrane resonators,” IEEE Trans.
Ultrason. Ferroelect. Freq. Contr., 51, pp. 1255-1261 (2004).

Efremov A, et al., “Etching mechanism of Pb(Zr, Ti)O3 thin films in Cl2/Ar plasma,”
Plasma Chemistry and Plasma Processing, Vol. 24, No.1, pp. 13-28 (2004)

Farlow R, Galbraith W, Knowles M, and Hayward G, “Micromachining of a
piezocomposite transducer using a copper vapor laser,” IEEE Trans. Ultrason. Ferroelec.
Freq. Control, Vol. 48, No. 3, pp. 639-640 (2001).

Foster FS, Pavlin CJ, Harasiewicz KA, Christopher DA and Turnbull DH,“Advances in
ultrasound biomicroscopy,” Ultrasound in Med.&Biol., vol 26, pp. 1-27 (2000).

99
Foster FS, Pavli CJ, Lockwood GR, Ryan LK, Harasiewicz KA and Rauth AM
“Principles and applications of ultrasound backscatter microscopy.” IEEE Trans.
Ultrason., Ferroelect., Freq. Contr, vol. 40, pp. 608-617 (1993).

Frood A, Beeby SP, Tudor MJ and White NM, “Photo patterned thick-film piezoelectric
elements on silicon,” J. Electroceramics., 19, pp. 327-331 (2007).

Gourverneur S, Lucat C, Menil F, Aucouturier JL, “New densification process of thick
films: Packaging,” IEEE Trans. Components, Packaging and Manufacturing Technology,
16, pp. 505-510 (1993).

Hu CH, Xu XC, Cannata JM, Yen JT and Shung KK, "Development of a real-time, high-
frequency ultrasound digital beamformer for high-frequency linear array transducers",
IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 53, pp.
317-323 (2006).

Huang O, Bandyopadhyay A, Bosa S, “Influence of processing parameters on PZT thick
films,” Materials Sciences and Engineering B., 116, pp.19-24 (2005).

Hunt JW, Arditi M and Foster FS, “Ultrasound transducer for pulse-echo medical
imaging,” IEEE Biomedical Engineering, vol 30, pp. 453-481 (1983).

Ioanna GM, Kim H, Kim I, Park SK, Choi K, et al, “High-frequency piezoelectric MEMS
ultrasound transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 54, pp.
2422-2430 (2007).

Jung J and Lee W, “Dry etching characteristics of Pb(Zr, Ti)O3 films in CF4 and
Cl2/CF4 inductively coupled plasmas,” Japan J. Appl. Phys., Vol. 40 Pt. 1, pp. 1408-
1419 (2001).

Kang M, et al., “Reduction of dry etching damage to PZT films etched with a Cl-based
plasma and the recovery behavior,” J. of the Korean Physical Society, Vol. 41, No. 4, pp.
445-450 (2002).

Kino GS, Acoustic Waves: Devices, Imaging, and Analog Signal Processing, New
Jersey, Prentice Hall, (1987).

Kokaze Y, et al., “Dry etching process for Pb(Zr, Ti)O3 thin-film actuators,” The Japan
Society of Appliced Physics, Vol. 46, No. 1, pp. 280-282 (2007).

Krimholtz R, Leedom DA, and Matthaei GL, “New equivalent circuit for elementary
piezoelectric transducers,” Electron. Letters, vol. 6, No. 13, pp. 398-399 (1970).

100
Kurchania R and Milne SJ, “Characterization of sol-gel Pb(Zr0.53Ti0.47)O3 films in the
thickness range 0.25-10 µm,” Journal of Materials Research.,14[5], pp.1852-1859
(1999).

Levassort F, Pascal L, Hue TH, Holc J, Bove T, Kosec M, and Lethiecq M, “High
performance piezoelectric films on substrates for high frequency imaging,” in Proc.
IEEE Ultrasonics Symposium, pp. 1035-1038 (2001).

Lockwood GR, Turnbull DH, Christopher DA and Foster FS, "Beyond 30MHz,
Application of High-frequency Ultrasound Imaging," IEEE Engineering in Medicine and
Biology, pp. 60-71 (1996).

Lukacs M, et al., “Laser micromachined high frequency ultrasonic arrays,” in Proc. IEEE
Ultrason. Symp., pp. 1209-1212 (1996).

Lukacs M, Olding T, Sayer M, Tasker R, Sherrit S, “Thickness mode material constants
of a supported piezoelectric film,” J. Appl. Phys., 85, pp. 2835-2843 (1999).

Lukacs M, Sayer M, Foster FS, “Single element high frequency (<50 MHz) PZT sol gel
composite ultrasound transducers,” IEEE Trans. Ultrason. Ferroelect. Freq. Contr., 47,
pp.148-159 (2000).

Maeda R, Wang Z, Chu J, Akedo J, Ichiki M, Yonekubo S, “Deposition and Patterning
Technique for Realization of Pb(Zr0.52Ti0.48)O3 Thick Film Micro Actuator,” Jpn. J.
Appl. Phys., 37 [12B], pp. 7116-7119 (1998).

Maki K, Soyam N, Mori S, and Ogi K, “Evaluation of CSD-PZT thick films with
different film density,” ISAF, 2, pp. 957-960 (2000).

Marechal P, Levassort F, Holc J, Tran-Huu-Hue L, Kosec M and Ldthiecq M, “ High-
frequency transducers based on integrated piezoelectric thick films for medical imaging,”
IEEE Trans. Ultrason., Ferroelect., Freq. Contr, vol 53, pp. 1524-1533 (2006).

Marks S, et al., “The profile and device characterization of high wafer temperature etched
Ir/PZT/Ir stacks,” Integrated Ferroelectrics, Vol. 59, pp. 333-340 (2003).

Mass R, Koch M, Harris NR, White NM, Evans AG, “Silicon micromachining processes
combined with thick-film printed lead zirconate titanate actuators for
microelectromechanical systems,” Materials Letters, 31, 109-112 (1997).

Mina IG, Kim H, Kim I, Park SK, Choi K, Jackson TN, Tutwiler RL, Trolier-McKinstry
S, “ High frequency piezoelectric MEMS ultrasound transducer,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr., vol. 54, pp. 2422-2430 (2007).

101
Morton CE, Lockwood GR, “Evaluation of kerfless linear arrays,” IEEE Ultrasonics
Symposium. pp. 1257-1260 (2002).

Muralt P, “Ferroelectric thin film for micro-sensor and actuators: a review,” J.Micromech
Microeng, Vol 10, pp. 136-146 (2000).

Navarro A, Alcock JR, Whatmore RW, “Aqueous colloidal processing and green sheet
properties of lead zirconate titanate (PZT) ceramics made by tape casting,” J. Euro.
Ceram. Soc., 24 [6], pp. 1073-1076 (2004).

Nelson SO, “Density-Permittivity relationships for powdered and granular materials,”
IEEE Trans. Inst. Meas., 54, pp. 2033-2040 (2005).

Pang GF, Sayer M, Lockwood GR and Watt M, “Fabrication of PZT sol gel composite
ultrasonic transducers using batch fabrication micromolding,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr, vol 53, pp. 1679-1684 (2006).

Rhee S, Ritter TA, Shung KK, Wang H, Cao W, "Materials for acoustic matching in
ultrasound transducers", Proceedings of the 2001 IEEE Ultrasonics Symposium, pp.
1051-1055 (2001).

Ritter T, Shung KK, Tutwiler R, Shrout T, "High frequency transducer arrays for medical
imaging", Proceedings of the SPIE - The International Society for Optical Engineering,
vol. 3982, pp. 208-212 (2000).

Robert AD, “Fabrication and characterization of annular thickness mode piezoelectric
ultrasonic transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 54, pp.
2462-2468 (2007).

Sakashita Y, Segawa H, Tominaga K and Okada M, “Dependence of electrical properties
on film thickness in Pb(ZrxTi1-x)O3 thin films produced by metalorganic chemical vapor
deposition," J. Appl. Phys., 73, pp.7857-7863 (1993).

Shepard JF, Moses PJ, and Trolier-McKinstry S, “The Wafer Flexure Technique for the
Determination of the Transverse Piezoelectric Coefficient (d31) of PZT Thin
Films,” Sens. Actuators., A 71 pp.133-138 (1998).

Sherar MD and Foster FS, “The design and fabrication of high frequency poly(vinylidene
fluoride) transducer,” Ultrasonic Imaging, vol 11, pp. 75-94 (1989).

Sherar MD, Starkoski BG, Taylor WB and Foster FS, “A 100 MHz B-scan ultrasound
backscatter microscope,” Ultrasonic imaging, vol 11, pp. 95-105 (1989).

Shung KK, Diagnostic Ultrasound and Blood Flow Measurements, Taylor & Francis Ltd,
(2005).

102
Shung KK, Zippuro M, “Ultrasonic transducers and arrays,” Engineering in Medicine
and Biology Magazine, vol.15, pp. 20-30 (1996).

Subasinghe S, Goyal A, and Tadigadapa S, “High aspect ratio plasma etching of bulk
lead zirconate titanate,” Proc. SPIE, Vol. 6109, 61090D (2006)

Sun L, Lien CL, Xu XC, Shung KK, "In vivo cardiac imaging of adult zebrafish using
high frequency ultrasound (45-75 MHz)", Ultrasound in Medicine and Biology, vol. 34,
pp. 31-39 (2008).

Takahiro Y, Masako K and Norikazu S, “Influence of poling conditions on the
piezoelectric properties of PZT ceramics,” Journal of Materials Science: Materials in
Electroncis., vol. 11, pp. 425-428 (2000).

Tanaka K, et al., “Fabrication of microdevices using bulk ceramics of lead zirconate
titanate,” The Japan Soc. Of Applied Physics, Vol 44, No. 98, pp. 7068-7071 (2005).

Trolier-McKinstry S, Muralt P, “Thin film piezoelectric for MEMS,” Journal of
Electroceramics., vol. 12, pp. 7-17 (2004).

Wang H, Ritter TA, Cao W, Shung KK, "High frequency properties of passive materials
for ultrasonic transducers", IEEE Transactions on Ultrasonics, Ferroelectrics, and
Frequency Control, vol. 48, pp. 78-84 (2001).

Wang HW, Cheng SY and Wang CM, “Optimization of poling process for piezoelectric
PZT ceramics,” IEEE/CHMT, Japan IEMT Symposium., Nara, Japan, April, pp. 263-266
(1989).

Wang S, et al., “Deep reactive ion etching of lead zirconate titanate using sulfur
hexafluoride gas,” J. Am. Ceram. Soc., 82(5) , pp. 1339-41 (1999).

Wang ZH, Zhu WG, Zhu H, Miao JM, Chao C, Zhao CL and Tan OK, “Fabrication and
characterization of piezoelectric micromachined ultrasonic transducers with thick
composite PZT films,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr, vol 52, pp.
2289-2297 (2005).

Wells PNT, “Ultrasonic imaging of the human body,” Rep. Prog. Phys. Vol 62, pp 671-
722 (1999).

Xu F, Wolf RA, Yoshimura T, and Trolier-McKinstry S, “Piezoelectric films for MEMS
applications,” 11th International Symposium on Electrets, pp. 387-396, (2002).

Xu YH, Ferroelectric Materials and Their Applications, North-Holland (1991).

103
Zhou QF, Cannata JM, Meyer RJ, Van Tol DJ, Hughes WJ, Shung KK, Trolier-
McKinstry S, “Fabrication and characterization of micro tonpilz high frequency
transducer derived by PZT thick films,” IEEE Trans. Ultrason. Ferroelect. Freq. Contr.,
52, pp. 350-357 (2005).

Zhou QF, Shung KK, Zhang Q and Djuth FT, "High Frequency Piezoelectric
Micromachined Ultrasonic Transducers for Imaging Applications," Proceedings of SPIE,
Kissimmee, Florida, April pp. 19-20 (2006).

Zhu W, Yao K, Zhang Z, “Design and Fabrication of a Novel Piezoelectric Multilayer
Actuator by Thick Film Screen Printing Technology,” Sensors and Actuators, 86, 149-
153 (2000).

Abstract (if available)

Abstract Fabrication of high-frequency (30 MHz -50 MHz) ultrasonic linear arrays is still a challenge. The task is even more difficult to build arrays at a frequency higher than 100 MHz, which has the potential to provide more detail skin texture for early diagnosis of melanoma or to image small objects such as stem cells. Integrating of PZT films into MEMS micro-machined is a potential solution to such high-frequency applications. This thesis presents the development of high-frequency (~100 MHz) PZT thick-film transducers and arrays.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A high frequency array- based photoacoustic microscopy imaging system

PDF

Development of high frequency focused transducers for single beam acoustic tweezers

PDF

Development of front-end circuits for high frequency ultrasound system

PDF

Array transducers for high frequency ultrasound imaging

PDF

High-frequency ultrasound array-based imaging system for biomedical applications

PDF

Design and development of ultrasonic array transducers for specialized applications

PDF

High frequency ultrasonic phased array system and its applications

PDF

High-frequency ultrasound imaging system with Doppler features for biomedical applications using 30~35 mHz linear arrays and an analog beamformer

PDF

2D ultrasonic transducer array’s design and fabrication with 3D printed interposer and applications

PDF

High frequency ultrasonic imaging for the development of breast biopsy needle with miniature ultrasonic transducer array

PDF

Fabrication of ultrasound transducer and 3D-prinitng ultrasonic device

PDF

Quantification of cellular properties using ultra-high frequency single-beam acoustic tweezer

PDF

Development of novel 1-3 piezocomposites for low-crosstalk high frequency ultrasound array transducers

PDF

High frequency ultrasound elastography and its biomedical applications

PDF

High-frequency ultrasonic transducers for photoacoustic applications

PDF

Novel beamforming with dual-layer array transducers for 3-D ultrasound imaging

PDF

Miniature phased-array transducer for colorectal tissue characterization during TEM robotic surgery; and, Forward-looking phased-array transducer for intravascular imaging

PDF

High frequency ultrasound array for ultrasound-guided breast biopsy

PDF

Microfluidic cell sorting with a high frequency ultrasound beam

PDF

Intravascular imaging on high-frequency ultrasound combined with optical modalities

Asset Metadata

Creator Wu, Dawei (author)

Core Title Development of high-frequency (~100mhz) PZT thick-film ultrasound transducers and arrays

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 09/10/2009

Defense Date 07/13/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag high-frequency,OAI-PMH Harvest,PZT film,ultrasonic

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Shung, K. Kirk (committee chair), Chen, Yong (committee member), Sadhal, Satwindar S. (committee member), Yen, Jesse T. (committee member), Zhou, Qifa (committee member)

Creator Email daweiwu@usc.edu,dwu.bme@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2595

Unique identifier UC1316239

Identifier etd-Wu-3193 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-254740 (legacy record id),usctheses-m2595 (legacy record id)

Legacy Identifier etd-Wu-3193.pdf

Dmrecord 254740

Document Type Dissertation

Rights Wu, Dawei

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu