University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

DEVELOPMENT OF NOVEL 1-3 PIEZOCOMPOSITES
FOR LOW-CROSSTALK HIGH FREQUENCY
ULTRASOUND ARRAY TRANSDUCERS

by
Hao-Chung Yang

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

August 2012

Copyright 2012 Hao-Chung Yang

ii

Acknowledgements

Throughout the course of the study, many important individuals have offered
advice and support that are invaluable to me, and it is my honor to receive their words of
wisdom. Prior to the presentation of this dissertation, I would like to give big credit to
these important individuals.
First, special thanks to my mother, Kuei-Ying Lee, and my father, Yung-Hsing
Yang, who have been so supportive of me throughout my life. I am so happy to have you
as parents. Also, many thanks to my elder sister, Ju-Min Yang, my younger sister, Shu-
Fei Yang, and my grandma, Kuo-Man Lee, for being there for me especially during the
last couple of years.
Next, I would like to express my sincere gratitude to my thesis advisor, Prof. K.
Kirk Shung, for giving me the freedom to explore and develop my own research
directions. Dr. Shung has always been very supportive of my work and has provided lots
of insightful comments to help me fine-tune my research results. Also, my lab manager,
Dr. Jonathan Cannata, has been an influential figure and mentor during my graduate
studies. Thanks for being so supportive of my education and for gladly taking the time to
explain transducers, piezoelectric materials, and everything else. I would also like to
thank my coworkers and friends, Jay Williams, Dr. Changhong Hu, Dr. Changgeng Liu,
Dr. Dawei Wu, Dr. Lisong Ai, Dr. Lequan Zhang, Fan Zheng, Rumin Chen, and Chi Tat
Chiu at USC for their help and encouragement while I finished up my studies. Finally, I
would like to thank the National Institutes of Health for providing the funding for this
research project through NIH grant P41-EB2182.

iii

Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vi
Abstract viii
Chapter 1: Ultrasonic Imaging and Transducers 1
1.1 Ultrasonic Imaging and High Frequency Application 1
1.2 Introduction of Single Element Transducers and Linear Arrays 7
1.2.1 KLM Model 14
1.2.2 Design Criteria for Single Element Transducers and Linear Array
Transducers 16
1.3 Scope of the Research 19
Chapter 2: Piezoelectric Materials and Composites 21
2.1 Piezoelectric Materials 21
2.2 Piezoelectric Composites 28
2.2.1 High Frequency Composites for Linear Arrays 28
2.2.2 Fabrication of High Frequency 1-3 Composites 34
Chapter 3: Composites Fabrication and Evaluation 39
3.1 Composites Design 39
3.1.1 Finite Element Modeling 42
3.2 Composites Fabrication 45
3.3 Composites Evaluation 48
Chapter 4: Composite Transducer Fabrication and Evaluation 52
4.1 Single Element Transducer 52
4.1.1 Single Element Transducer Fabrication 52
4.1.2 Single Element Transducer Evaluation 53
4.2 Array Transducer Sub-aperture 58
4.2.1 Array Transducer Sub-aperture Fabrication 58
4.2.2 Array Transducer Sub-aperture Evaluation 60
4.2.2.1 Pulse-echo Measurement 60

iv

4.2.2.2 Insertion Loss Measurement 62
4.2.2.3 Crosstalk Measurement 64
Chapter 5: High-Frequency Kerfless Annular Array Transducer 67
5.1 Annular Array Transducer 67
5.2 Array Transducer Design and Fabrication 69
5.3 Array Transducer Evaluation 73
5.4 Array Transducer Imaging 77
Chapter 6: Summary and Future Work 82
Bibliography 85

v

List of Tables

Table 2.1 Parameter ranges for piezoelectric ceramics, polymers, and composites. 27

Table 3.1 Designed parameters for the PZT-5H 1-3 composites with different pillar
geometries for this study. 40

Table 3.2 Average measured material properties for the PZT-5H 1-3 composites with
different pillar geometries manufactured for this study. 51

Table 4.1 Measured frequency response properties for the PZT-5H 1-3 composites
transducers with different pillar geometries. 57

Table 4.2 Measured insertion lose (IL) and – 20 dB pulse length (PL) for the 15
MHz array sub-apertures with different pillar geometries. Element 1 to
8 of each the array sub-apertures were measured to represent all the
elements. 61

Table 4.3 Average measured the combined crosstalk (electrical and acoustical) for
the 15 MHz array sub-apertures with different pillar geometries. Element
1 to 8 of each the array sub-apertures were measured to represent all the
elements. 66

Table 5.1 Specifications for the 40 MHz 1–3 composite annular array transducer. 69

Table 5.2 The measured pulse echo characteristics for all annular array elements. 75

vi

List of Figures

Figure 1.1 (a) Projected growth of diagnostic imaging market from 2010 to 2016.
(b) Subgroups comprising the diagnostic imaging market in 2010. 2

Figure 1.2 Illustration of the formation of A-mode and B-mode image displays. 6

Figure 1.3 A cross-sectional drawing of the single element ultrasound transducers. 9

Figure 1.4 A cross-section of a linear array ultrasound transducer. 12

Figure 1.5 KLM electrical equivalent model. 16

Figure 2.1 Piezoelectric geometries and associated electromechanical coupling
coefficient for single element transducers, 2-2 composites and 1-3
composites transducers. 30

Figure 2.2 Two different configurations of piezoceramic composites:
(a) 1–3 composite; (b) 2–2 composite. 31

Figure 3.1 3D geometry of the composites with different pillars constructed
by PZFLEX. 43

Figure 3.2 Modeled electric impedance and phase angle of the composites with
different pillars using PZFLEX 3D model. 44

Figure 3.3 Modeled electric impedance and phase angle of the composites with
different pillars using PZFLEX 3D model. 47

Figure 3.4 Measured electrical impedance (solid line) and phase angle (dashed line)
for the piezocomposites with different pillar geometries. At 10 MHz,
square (a), triangle (b), and pseudo-random (c). At 15 MHz. square (d),
triangle (e), and pseudo-random (f). 49

Figure 4.1 A cross-sectional drawing of the single element ultrasound transducers. 53

Figure 4.2 Block diagram of a pulse-echo measurement, which is the most common
test for transducers. 55

Figure 4.3 Measured pulse-echo waveform (a), their spectrums (b), and envelopes
(c) at 10 MHz for the composite single element ultrasound transducers.
Measured pulse-echo waveform (d), their spectrums (e), and envelopes
(f) at 15 MHz for the composite single element ultrasound transducers. 56

vii

Figure 4.4 Sputtered 1-3 composites with different pillar geometries (removing the
Cr/Au over the epoxy kerfs) (a) Square (b) 45° Triangle
(c) Pseudo-Random. 59

Figure 4.5 A picture of the completed array sub-aperture. 59

Figure 4.6 Block diagram of an insertion loss measurement, which is the most
common test to evaluate the sensitivity of a transducer. 63

Figure 4.7 Block diagram of a cross-talk measurement for array elements. 65

Figure 4.8 Average measured the combined electrical and acoustical crosstalk
between the nearest and next-nearest elements of the array sub-apertures
with different pillar geometries. 66

Figure 5.1 AutoCAD drawing of 6-element flexible circuit and array elements. 71

Figure 5.2 The diced PR composites with it flexible circuit. Epoxy 301 was used to
glue the backed composite to the assembly. 71

Figure 5.3 A schematic section drawing of the annular array showing all major
components. 72

Figure 5.4 A picture of the completed 40 MHz annular array transducer. 72

Figure 5.5 Measured time domain pulse echo response (solid line) and normalized
frequency spectrum (dashed line) for element #3. 75

Figure 5.6 Cross-talk measured between adjacent elements of the kerfless annular
array. 76

Figure 5.7 Measured wire phantom image with a 60 dB dynamic range. 80

Figure 5.8 Measured cyst phantom image with a 50 dB dynamic range. 81

viii

Abstract
The goal of this research was to develop a novel diced 1-3 piezocomposite
geometry to reduce pulse-echo ring down and acoustic cross-talk between high-frequency
ultrasonic array elements. PZT-5H based 1-3 composites (10 and 15 MHz) of different
pillar geometries (square (SQ), 45° triangle (TR), and pseudo-random (PR)) were
fabricated and then made into single element ultrasound transducers. The measured
pulse-echo waveforms and their envelopes indicate that the PR composites had the
shortest -20 dB pulse length and highest sensitivity among the composites evaluated.
Using these composites, 15 MHz array sub-apertures (0.95 λ pitch) were fabricated to
assess the acoustic cross-talk between array elements. The combined electrical and
acoustical cross-talk between the nearest array elements of the PR array sub-apertures (-
31.8 dB @ 15 MHz) was 6.5 dB and 2.2 dB lower than that of the SQ and the TR array
sub-apertures, respectively. To further reduce inter-element cross-talk of kerfless annular
array transducers, utilizing the piezocomposites with the PR pillar geometry, we
fabricated high-frequency annular arrays. Each annular array was designed to have six
equal-area elements and a center frequency of 40 MHz. The average center frequency
estimated from the measured pulse-echo responses of array elements was 38.7 MHz and
the −6 dB bandwidth was 51%. The average insertion loss recorded was 23.1 dB, and the
maximum combined crosstalk between the adjacent elements was less than −31 dB.
These results demonstrate that the 1-3 piezocomposite with the pseudo-random pillars
may be a better choice for fabricating enhanced high frequency linear array ultrasound
transducers; especially when mechanical dicing is used.

1

Chapter 1: Ultrasonic Imaging and Transducers
This chapter, emphasizing the design of transducers and arrays, discusses basic ultrasonic
imaging concepts. A- and B-mode imaging is each explained to give basic knowledge about
image acquisition and display. The importance of the ultrasonic transducer for ultrasonic
imaging is then discussed with emphasis on single element transducers and linear array
transducers.

1.1 Ultrasonic Imaging and High Frequency Application
The global diagnostic imaging market is rapidly growing due to the increased awareness of
the importance of the early screening for disease. This market is expected to continue to grow
from $20.7 billion (in 2010) to $26.6 billion (by 2016). Ultrasonic imaging is one of the major
imaging modalities used for medical diagnosis, with the others being X-ray, including computed
tomography (CT), magnetic resonance imaging (MRI), and nuclear imaging. Among them,
ultrasound (21%, 2010) is one of the most widely used (being only second to X-ray, 34%, 2010)
(see Figure 1.1). This rapidly growing imaging modality is extremely popular for several reasons.
First, ultrasonic scanners are considerably cheaper and faster than both CT and MRI scanners.
Second, for moving tissues, such as in the heart, real time imaging is required for diagnosis,
which ultrasound also fulfills this need. Third, ultrasound is the only non-invasive imaging
modality at the power level used by imaging scanners. Fourth, ultrasound is capable of obtaining
other physiological data such as blood flow velocity and pressure. In addition, ultrasound
machines are typically more portable and light-weight than other imaging modalities.

2

Figure 1.1 (a) Projected growth of diagnostic imaging market from 2010 to 2016. (b) Subgroups
comprising the diagnostic imaging market in 2010.

The process of forming an ultrasound image is simple to understand but quite
complicated to perform. Ultrasound waves are generated and detected by transducers that operate
similarly to a speaker and microphone combination. When examining tissues, a transducer sends
ultrasonic wave out and receives its reflecting from the tissues internal structures. The delay in
time and amplitude of a reflected echo returning to the transducer is measured using modern
electronic instruments. With knowledge of the speed of sound in the tissue, the distance to the
reflecting structure can be calculated and displayed. This is the core idea of forming an image for
the majority of ultrasound imaging techniques.
In ultrasonic imaging, more than a few principal echo display modes are used. Figure 1.1
shows two of the most common modes. The A-mode, or amplitude mode, displays the instant
echo amplitude versus the time after transmission. The time between reflections can easily be
converted to distance by assuming a certain speed of sound. Although the speed of sound in
biological tissues varies, we typically assume that the propagation speed of sound is 1540 m/s.
Once received, the acoustic echo is transformed to an RF signal by the transducer, where it is

3

recorded by the imaging system electronics. The measured echo signal is then converted into A-
mode signal through a process of rectifying and smoothing. This process, as known as signal
demodulation, converts the echo signal burst to a single spike for each reflector. At the present
time, the A-mode ultrasound scanners have been adopted for many clinical applications, such as
in the ophthalmologic industry, the most common medical application for A-mode imaging. For
instance, an ultrasonic pachymeter is able to quickly measure thickness of various structures in
the ocular cavity such as the cornea.
The other imaging mode is B-mode, or brightness mode. In this mode, echo signals are
converted electronically to intensity modulated dots on the screen. The brightness of the dot is
proportional to the echo signal amplitude. (Figure 1.2)
An ultrasonic image is constructed from echoes returned from reflecting or scattering
targets. The time that elapses between the transmitted pulses and received echoes provides the
positional information of the image; and the brightness of the image depends on the amplitude of
the echoes. Another important aspect of an ultrasonic image is its spatial resolution since that
will determine how small an object can be differentiated from the surrounding structures. The
spatial resolution depends on the profiles of the ultrasonic beam and pulse excited by a
transducer. Axial spatial resolution (in depth) is determined by the duration of the ultrasound
pulse within -6-dB of the maxima. 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. In theory, for a circular focused transducer, the axial resolution, R
axial
, and the lateral
resolution, R
lateral
at the focal point can be estimated using the following equations:
axial
c
c
R=
2 BW F 
(1.1)

4

where c is the velocity at the loading medium; F
c
is the center frequency of a transducer; and
BW is the -6 dB bandwidth of the transducer.
lateral
c
ff
R = λ
d F d
c 


(1.2)
where f is the focal distance; d is the diameter of the aperture; and λ is the wavelength in the
medium. Clearly, as the frequency increases, the spatial resolution improves. Therefore, the
spatial resolution of an ultrasound image is basically determined by the operating frequency of
the transducer.
This is especially important for medical imaging applications, such as imaging of human
skin, eyes and vascular anatomy and small animals. Since these applications require finer spatial
image resolution; however, conventional clinical ultrasound imaging systems, which are
operated at between 1 MHz and 10 MHz, lack of ability to fulfill this need. To improve this
spatial resolution, ultrasound imaging systems have to be operated at a higher frequency. This
need has driven researchers in the medical ultrasonic industry to develop higher frequency
imaging systems. In the past 20 years, numerous high frequency (>20 MHz) ultrasound were
made for a variety of imaging applications, such as dermatology (Passman et al., 1993; Turnbull
et al., 1995), ophthalmology (Silverman et al., 2001), and small animal studies (Turnbull et al.,
1999; Foster et al., 2002). In 2002, Foster pushed the limit of the operating frequency beyond 30
MHz by demonstrating assessment of cardiac structures and functions in mouse embryos using
an ultrasound backscatter microscopy (abbreviated as UBM). Since then, the ultrasound
microscopy not only provided much finer imaging resolution for researchers, but also
encouraged them to investigate new research areas.

5

However, the advantages of increased device resolution come at the expense of signal
penetration depth since attenuative losses observed in various human soft tissues increase as the
propagating sound wave frequency increases (Shung et al.,1992; Foster et al., 2000). In other
words, the increased signal loss at high frequencies limits the depth of penetration and the
resolution of the ultrasound beam. To overcome these drawbacks, it is important to optimize the
sensitivity and bandwidth of the imaging system accordingly since high frequency ultrasonic
imaging brings challenges not only in design of imaging systems, but also in the development of
ultrasonic transducers. In our group, for high frequency ultrasound imaging applications,
Cannata et al. (2004) have developed a 64-element composite array transducer operating at 35
MHz. To pair with this array transducer, in 2006, Hu et al. have developed a high frequency
linear array ultrasonic imaging system. However, for the ultrasound transducers, there is still
plenty of room for improvement, such as stronger sensitivity, wider bandwidth, and lower cross-
talk. Therefore, in this research, to obtain enhanced quality of ultrasound imaging, we are
interested in designing and fabricating the ultrasonic transducers operating at high frequencies.

6

Figure 1.2 Illustration of the formation of A-mode and B-mode image displays, from K. K.
Shung (2006)

7

1.2 Introduction of Single Element Transducers and Linear Arrays
A transducer is a device that coverts one form of energy to another, such as from
mechanical energy to electrical energy, and vice versa. Figure 1.3 shows a cross-section of a
typical single element transducer with major parts labeled. A single element transducer
incorporates five major parts: a piezoelectric element, match layers, an acoustic backing layer, an
electrical impedance matching network, and a lens for focusing the sound beam.
A piezoelectric element is a resonant device; and its thickness determines the resonant
frequency. As previously noted, the piezoelectric element of an ultrasound transducer converts
electrical energy to acoustical energy and vice versa. In order to increase the conversion
efficiency between the two energy forms, the impedance of the piezoelectric element must be
electrically matched to the electronics driving the element, as well as, acoustically matched to the
front load medium, such as human tissue. To produce better spatial resolution, the emitted sound
wave must also be focused, and a short time-domain response must be achieved. To apply an
electric field to the element, metallization layers on both faces are required. The negative
electrode layer is often connected to a metal transducer housing which is attached to an electrical
adapter, provides RF shielding to the device. For ultrasonic transducers, the most commonly
used piezoelectric materials are lead zirconate titanate (PZT) ceramics, single crystals such as
lithium niobate (LiNbO3), and piezoelectric polymers such as polyvinylidene fluoride (PVDF)
and its copolymer. These piezoelectric materials are discussed in Chapter 2.
Acoustic matching layers are used to improve the performance of a transducer. DeSilets
et al., 1978 demonstrated that a single matching layer should be equal to 0.25 λ and its acoustical
impedance (Z
m
) can be estimated using the following equation:

8

 
1
2 3
12 m
Z Z Z  (1.3)
where Z
1
is the impedance of the piezoelectric element, and Z
2
is the impedance of the load
medium. To fabricate the matching layers, composite materials consisting of a polymer and
small particles of silver, graphite, and tungsten are commonly used.
The backing material has two major functions. It provides a rigid support for the fragile
piezoelectric element and enhances the emitted sound waves. As the acoustic impedance of the
backing layer approaches that of the piezoelectric element, the backing layer dampens element
vibrations and decreases the mechanical quality factor (Q) of a device. These are able to reduce
rign-down time. Therefore, by producing a shorter time domain response, a heavy backing is
able to improve image resolution. However, because the heavy backing leads to more mass
loading, this improvement comes at the cost of reduced device sensitivity and frequency. The
other important function of the backing material is signal absorption. The absorption of
mechanical signals transmitted into the backing layer must be optimized in order to avoid image
artifacts caused by backing echoes. A backing layer can also be used to dissipate heat and
provide electrical connections to the piezoelectric element.

9

Figure 1.3 A cross-sectional drawing of the single element ultrasound transducers.
An electronic matching circuit is used to match the electric impedance of the
piezoelectric element to the transmit and the receive electronics in order to enhance the
efficiency of electrical energy transmission. In general, both on the transmit and the receive
circuits, most of the transducers are coupled with an electronic load of 50 Ω. To do this, the
dielectric permittivity, thickness, and area of a piezoelectric element must be carefully selected
to match the real part of the impedance (resistance) to 50 Ω; series and shunt inductors can be
used to cancel out the imaginary part of the impedance (reactance) at the device center frequency
as well. If the impedance still does not reach 50 Ω, in addition to these steps, a transformer can
then be used to further match the real part of the impedance to 50 Ω. Since most of ultrasonic

10

transducers require a broad bandwidth response, the electrical matching element should have to
be effective over the device pass-band.
For a flat unfocused transducer, the width of the emitted ultrasound beam at the focal
point (at the near-field/far-field transition point) is typically too wide to provide the lateral
resolution needed for most imaging applications (Shung et al., 1992). Therefore, to improve the
lateral imaging resolution, acoustic focusing is highly preferred. Typically, either shaping the
piezoelectric element or adding an acoustic lens is used to obtain a narrower beam width at the
focus. However, this improvement in the lateral resolution comes at the cost of a limited depth of
field. As with the matching layers, the acoustic impedance of the lens material should be well-
matched to the impedance of the load medium. Materials with low acoustic attenuation, such as
Polymethylpentene (TPX), are also desired. Concave or convex lenses can be used for this
operation. Using Snell’s Law the type of lens formed is dependent on the sound velocity of the
material used. For materials displaying a sound velocity less than that of the load medium, a
convex lens is required for focusing. Conversely, for materials with a higher sound velocity than
the load medium a concave lens is used. Most commercial linear arrays incorporate convex
lenses made out of silicone or urethane polymers. These polymers are not typically used for high
frequency arrays due to increased signal attenuation. At high frequencies lenses are typically
concave and made from harder materials such as unloaded epoxy (Cannata et al., 2003),
polystyrene, quartz, or TPX (Ritter, 2000).
Ultrasonic array transducers have the same basic structure as the single-element
transducer described above, except that each array element is designed to act as a single isolated
transceiver. These elements may be shaped like a ring and arranged concentrically (called
annular array); like a rectangle and arranged in a line (called linear or one-dimensional array); or

11

like a square and arranged in rows and columns (called two-dimensional array). Compared to
single element transducers, annular arrays provide excellent image quality but still require
mechanical scanning to form an ultrasonic image. To avoid the scanning, one-dimensional arrays
offer significant improvements in performance. On the other hand, multi-dimensional arrays
have the advantage of improved image quality due to the use of dynamic focusing and beam
steering in multiple planes across the volume of interest. However, because fabricating and
implementing these devices (especially for multi-dimensional arrays) is very difficult, the
multidimensional arrays are limited to less than 40 MHz at the present time.
Figure 1.4 is a representation of a typical one-dimensional array structure. The
construction of a linear array is very similar to that of a single element transducer. Linear arrays
incorporate electrical impedance networks, acoustic matching and backing layers, and employ
some method for mechanically focusing the sound beam in the dimension that is not focused
electronically.

12

Figure 1.4 A cross-section of a linear array ultrasound transducer, from J. Cannata (2004)

13

To produce a two-dimensional ultrasound image, the ultrasound beam is necessary to be
scanned across the imaging plane either mechanically by a single element transducer or
electronically by an array. To scan the target mechanically it only requires a simple imaging
system and transducer design, but mechanical scanning normally needs very bulky positioning
equipment to control and scan the transducer. Therefore, because of the transducer vibration and
fixed focus zone, the image quality of the mechanical scanning is inferior to that of electronically
scanning. Linear array scanning, on the other hand, allows the ultrasonic beam to be steered and
focused electronically, which provides high-quality real-time images. However, the fabrication
of linear arrays is more complicated, and the related imaging systems are more expensive. In all,
both the single-element transducers and the linear arrays share the same criteria as a good
medical transducer: an ideal transducer for ultrasonic imaging should have high transmitting
efficiency and receiving sensitivity, wide dynamic range and frequency response, and a well-
matched impedance to the loading medium (usually human tissue).
To fulfill these criteria, before fabrication both mechanical matching and electric
matching must be carefully designed and modeled. Several 1-D transducer models such as the
Mason, the Redwood, and the KLM have been proposed and used. Among them, the KLM is the
most widely used since it is physically intuitive. It provides useful starting point for optimizing
the performance of transducers; design considerations of the single-element transducer and the
linear array are then discussed.

14

1.2.1 KLM Model
In 1970, Krimholtz et al introduced a one-dimensional design model as known as KLM
model. 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. 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. (see Figure 1.5)
The components and constants 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
Z
c
= radiation impedance of the piezoelectric layer (=ρcA)
ε
*
= complex clamped dielectric permittivity
e
33
= piezoelectric constant
c
D
= elastic constant of piezoelectric layer
ε
r
S
= clamped dielectric constant
0
A
C
d


 (1.3)
2
2 33
t DS
r
e
k
c 
 (1.4)
0
00
2
,
()
t
c
k c
C Z d
 



   (1.5)

15

' 0
21
0
(sin ( ))
t
C
C
kc



 (1.6)
Now that the model is in place, we can determine the input impedance of the
piezoelectric transducer in terms of the model parameters, the electric impedance of the
transducer (the input impedance of the KLM model) is given as
12
'2
12
11
( ( ))
in
o
ZZ
Z
j C j C Z Z 
  

(1.7)
where Z1 and Z2 are the respective radiation impedances of the medium into which the crystal is
radiating, respectively. A detailed discussion of acoustical impedances can be found in [Kinsler
et al., 2000]. With the KLM model, equations that describe insertion-loss and pulse-echo
response can also be obtained. From design point of view, the KLM model allows an intuitive
approach to optimize the transducer performance. Based on this model, several popular
transducer simulation software packages have been used to model the performance of the
transducers, such as PiezoCAD. It uses a chain matrix technique to calculate the overall system
characteristics from the electric terminals to the front acoustic port. This software also provides a
selection of piezoelectric material parameters from extensive piezoelectric database tables,
including plate, beam, and bar mode elements in ceramic, crystalline, polymer, and composite
materials. Therefore, PiezoCAD was extensively used in the project to quantitatively assess the
transducer response to various transmitted voltage or received pressure waveforms.

16

Figure 1.5 The KLM electrical equivalent model (D. Wu, 2010)

1.2.2 Design Criteria for Single Element Transducers and Linear Array Transducers
For designing single element transducers, several simple criteria must be followed. The
first rule is that the thickness of the piezoelectric material. If we consider the two surfaces of the
piezoelectric element as two independent vibrators, the resonant frequencies for such a
transducer are given by the following equation, (with the lowest resonant frequency being n = 1)
0
2
p
nc
f
L
 (1.9)
where C
p
is the acoustic wave velocity in the transducer material; L is the thickness of the
piezoelectric material; n is an odd integer. This equation indicates that resonance occurs when L
is equal to an odd multiple of a half wavelength.

17

The second rule is that, for a single element piston transducer with a radius (a) and
wavelength ( λ) in the loading medium, the transducer’s focus distance (the last maximum of
the axial pressure) was determined by
2
0
a
Z

 (1.8)
The region between the transducer surface and focus is called the Fresnel zone (near-field zone),
where the axial pressure oscillates greatly. On the other side of the Fresnel zone, the region
beyond the distance, the Far-field zone, is the ultrasound imaging region since in this zone the
axial pressure decreases gradually. This means the target of interest should be always located
beyond Z
0
.
In addition to these criteria, the array transducers, which consist of a large number of
identical transducer elements shaped like a rectangle in a linear arrangement, require extra rules.
If we assume that the width of the element is b and the space between any two elements, a kerf,
is d; the pitch, g, of this array will be the sum of b and d. The kerfs may be filled with acoustic
isolating material to decrease acoustic cross-talk and provide rigid support for array elements.
The radiation pattern of a linear array in the far field is represented by
 
1
sin sin
N
a
m
Lu bu
H u c u m c
g





 
   
  
 


(1.9)
where u = sinφ x, H(u) is the direction function at the angle of φ x, L
a
is the length of the array, N is
the number of the elements, b is the width of the element, g is the pitch, which is the space
between the center of two adjacent elements, and λ is the wavelength in the loading medium.
This equation shows that the magnitude of the grating lobes relative to the main lobe is
determined by the element width (b). As the element width decreases, the magnitude of the

18

grating lobes relative to the main lobe increases. In addition, the width of the array, L
a
, also
determines the width of the main lobe. The size of the main lobe decreases as the size of L
a

increases.
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:
1
sin
g
n
g






(1.10)
For the grating lobes to occur at angles greater than 90°, g must be smaller than λ/2. To remove
the grating lobes of the array, this condition must be satisfied.

Shung (2006) also summarized a few simple design rules for linear arrays. The pitch, g,
must be between 0.75 and 2λ; the ratio of the width of the array element to the thickness of the
element, b/L must be less than 0.6; to avoid spurious resonant modes, b must be greater than λ/2;
and the cross-talk between adjacent elements must be less than - 30 dB. These criteria for
designing ultrasound transducers are extensively used in this study.

19

1.3 Scope of the Research
This paper describes the design, fabrication and characterization of 1-3 piezocomposites
with pseudo-random pillars, 15 MHz kerfless linear array sub-apertures (16 elements), and 40
MHz kerfless annular arrays (6 elements). The goal of this research is to demonstrate that a novel
1-3 piezocomposite geometry can be successfully fabricated using the dice-and-fill technique,
and this geometry is able to reduce pulse-echo ring down and acoustic cross-talk between high-
frequency ultrasonic array elements.
Chapter 1, emphasizing the design of transducers and arrays, discusses basic ultrasonic
imaging concepts. The importance of the ultrasonic transducer for ultrasonic imaging is then
discussed with an emphasis on single element transducers and linear array transducers. Chapter 2
discusses the effect of properties of piezoelectric materials on ultrasound transducer design.
Important ferroelectric and piezoelectric properties of the piezoelectric materials are briefly
explained. Included is a description of a method for fabricating high frequency composites using
mechanical dicing. Chapter 3 describes the design and fabrication of the pseudo-random 1-3
piezocomposites. It also summarized the results, which show the reasons why the pseudo-
random pillar geometry is superior to other pillar geometries.
In Chapter 4, we provide the modeling, fabrication and evaluating of the pseudo-random
1-3 piezocomposite single-element transducer and the linear array sub-apertures. The
performance of the composites transducers with the pseudo-random pillars is characterized and
compared with those with the square or triangle pillars. In Chapter 5, we demonstrate the
development of a high-frequency kerfless annular array with the pseudo-random 1-3
piezocomposites. At the end of this chapter, the acquisition of wire phantom imaging using the

20

kerfless annular array transducer is presented. Chapter 6 includes the summary of the results,
conclusions and recommendations for future work.

21

Chapter 2: Piezoelectric Materials and Composites
This chapter discusses the properties of piezoelectric materials used in ultrasound
transducers. We compare the properties of piezoelectric ceramics, piezoelectric polymers, and
piezoelectric composites, and then detail the effect of their properties on ultrasound transducer
design. The fundamental equations and parameters governing the properties of the piezoelectric
materials are described. Included is a description of a method for fabricating high frequency
composites using mechanical dicing. The relationship between resonant frequency and
piezoelectric composite design considerations are also outlined.

2.1 Piezoelectric Materials
Piezoelectric materials, which convert electrical energy into mechanical when generating
an ultrasonic pulse and then convert mechanical energy back into an electric signal once it
detects the echoes. Piezoelectric materials determine the performance of ultrasonic transducers.
There are quite a number of piezoelectric materials to choose from when designing a single
element or array transducer. To narrow the choices, one must look at the piezoelectric properties
of these materials in the orientation of interest. In this section, we will compare several popular
piezoelectric materials: piezoelectric crystals (ex: Quartz (SiO
2
)), piezoelectric ceramic (ex:
Barium titanate (BaTiO3), Lead zirconate titanate (PZT)), piezoelectric polymers (ex:
Polyvinylidence difluoride (PVDF) and its copolymers), and piezoelectric composites (ex: PZT +
epoxy).
Many crystals (natural and synthetic) have piezoelectricity. One of the most well-known
naturally occurring piezoelectric crystals is quartz. In 1880, Jacques and Pierre Curie discovered
quartz's piezoelectric properties. However, the piezoelectricity of the naturally occurring

22

piezoelectric crystals is typical low, such as electromechanical coupling coefficient (k) of quartz
is only 0.09. This is the reason why today a common piezoelectric use of quartz is as a crystal
oscillator, such as the quartz clock. To increase piezoelectricity, therefore, many synthetic
piezoelectric crystals have been made, such as lead magnesium niobate-lead titanate (PMN-PT)
and lead zinc niobate-lead titanate (PZN-PT), which are relaxor-based ferroelectric single
crystals. They both have much higher electromechanical coupling coefficients of around 0.9.
Their superior properties make them are very suitable for fabricating high-sensitivity broad
bandwidth medical imaging transducers. Since these single crystals do not have grain size
consideration, they are preferred to be used in high frequency applications. However, one of the
drawbacks of the piezoelectric single crystals is their low curie temperatures (150 º C), which
render them unsuitable for high temperate applications.
Piezoelectric ceramics are created by mixing piezoelectric powder with a binder,
followed by pressing at high pressure and firing at high temperature. Piezoelectric ceramics
obtain very high electromechanical coupling coefficients, a broad range of dielectric constants
and very low losses. These advantages have made them the most widely-used materials for
ultrasonic imaging transducers. The ceramics are considered to be approximately isotropic since
their domains are randomly orientated. The randomly orientated domains largely decrease their
piezoelectric properties. To increase the piezoelectric properties, therefore, a poling process is
required to align the domains in the same direction. The poling process is applying a DC electric
field at a temperature close to Curie point (100 - 200 °C) for couple minutes, and then the
temperature is slowly decreased to room temperature while still applying the electric field.
Barium titanate (BaTiO
3
) and lead zirconate titanate (PZT) are the two most popular
piezoelectric ceramics for medical ultrasound imaging applications. Compared with BaTiO
3
,

23

PZT ceramics have superior piezoelectric properties, making them the most popular piezoelectric
materials for medical transducers.
Another kind of popular material for medical ultrasonic transducers is piezoelectric
polymers, such as Polyvinylidence difluoride (PVDF). By stretching the film and applying an
electric field of 300 kV/cm at 100 °C, the piezoelectric property of PVDF was firstly discovered
by Kawai, in 1969. Since then, PVDF gets a good reaction from researchers due to the fact that it
has many advantages over piezoelectric ceramics and crystals. For example, since its acoustic
impedance (~ 4 MRayls) is close to that of human tissues (~ 2 MRayls), acoustic matching layers
are not required to obtain a broad bandwidth of the transducer. In addition, it is quite flexible,
allowing press-focused. The 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. It is an excellent piezoelectric material for high frequency ultrasound
receivers, such as hydrophones, because of its close impedance match to water and high
receiving constant. However, its low electromechanical coupling, very low dielectric constants,
and high dielectric loss have limited their uses in many applications, although the recent
development of P(VDF-TrFE) co-polymers has shown a higher electromechanical coupling
coefficient.
Because the piezoelectric properties of piezoelectric materials vary, no piezoelectric
material is suitable for all transducers. For example, materials such as lithium niobate (LiNbO
3
),
lead titanate (PbTiO
3
), and PVDF copolymer are excellent choices for fabricating large aperture
single element transducers due to their low dielectric constants. In contrast, a material such as
PZT-5H (Navy Type VI) is ideal for array elements which typically have much smaller element
areas. There are still some other piezoelectric properties play important roles in determining the

24

performance of the transducers. For example, the grain size of piezoelectric materials is crucial
for the transducers operating at high frequencies. Piezoelectric materials used for high frequency
applications typically require a small size of grain to avoid the effect of grain boundaries.
Therefore, in recent years, researchers have been directed to the development of fine grain
piezoelectric ceramics (Hackenberger et al., 1996; Yu et al., 2002) and single crystal
piezoelectric materials (Park et al., 1997; Oakley, 2000; Ritter et al., 2000).
Since the properties of the piezoelectric materials are quite important, we detail the basic
parameters of the piezoelectric properties next. In this study, several assumptions are used to
simplify the fundamental constitutive relationships for an array element. First, the material is
assumed to be a uni-axial (isotropic in the plane perpendicular to the poling axis). Because of
this symmetry, the performance of a piezoelectric material can be completely described by five
elastic constants, three piezoelectric constants, and two dielectric constants. Next, we assumed
that no shear strains act on this element (no longitudinal stresses in the x-direction). Thirdly, the
length of the element in the y-direction is assumed to be infinite (no longitudinal strains in y-
direction). Therefore, the propagation velocity (V
33
’) and coupling coefficient (k
33
’) for a narrow
strip resonator can be calculated by first examining the dielectric permittivity (ε
33
’), elastic
stiffness constant (C
33
’) and piezoelectric stress constant (e
33
’). The full derivation of these
constants can be found in the works of DeSilets, 1978 and Selfridge, 1983. The dielectric
permittivity can be determined from the following relationship:
2
31
33 33
11
'
SS
E
e
C
  (2.1)

25

The quantity
2
31
11
E
e
C
is typically less than 5% of
33
S
 and is often ignored, leading to the
approximation
33 33
'
SS
  (Ritter, 2000). The elastic stiffness under constant electric field,
33
'
E
C , is
defined as:
 
2
13
33 33
11 33
'1
E
EE
EE
C
CC
CC





(2.2)
and piezoelectric stress constant ,
33
' e , is:
13 13
33 33
11
'
E
E
eC
ee
C
 (2.3)
The velocity of propagation under a constant electric field,
33
' V , is found by the following
formula:
2
33
33
33 33
33
( ')
'1
''
'
E
ES
e
C
C
V






 (2.4)
where ρ is the density of the materials. This equation can be used to determine the half
wavelength parallel resonant frequency,
p
f , using the following formula for an element of
height H:
2
33
33
33 33
( ')
'1
''
1
2
E
ES
p
e
C
C
f
H






 (2.5)

26

The general electromechanical coupling coefficient, k, provides a measure of the energy
conversion process in a piezoelectric material and is defined by Berlincourt in 1971 using the
following relationship:
2 delivered
delivered stored
W
k
WW


(2.6)
where W
delivered
is the energy delivered to the load medium per cycle and W
stored
is the energy
stored in the piezoelectric material per cycle. The coupling coefficient is a useful indication of
the efficiency of a piezoelectric material in a given geometry. Unlike other measures of
piezoelectricity, such as the strain constant (d
ij
) and the stress constant (e
ij
), a high coupling
coefficient points to high device sensitivity and bandwidth (Ritter, 2000). The coupling
coefficient for an array element, k
33
’, is described by following formula in terms of previously
defined properties:
2
33
33 33
33
2
33
33 33
( ')
''
'
( ')
1
''
ES
ES
e
C
k
e
C







(2.7)
This coupling coefficient is always greater than the thickness mode coupling coefficient,
k
t
, for a thin disk resonator. This is because the array element is clamped in only one lateral
direction. In contrast, k
33
’ is less than the coupling coefficient for a tall, narrow rod, k
33
, which is
not constrained in either lateral direction. In Table.2.1, important properties of the piezoelectric
ceramics, polymers, and composites are listed. This study focuses on the properties of the
piezoelectric composites.

27

Table 2.1 Parameter ranges for piezoelectric ceramics, polymers, and composites

Parameter ranges for piezoelectric ceramics, polymers, and composites
Material Parameter Piezoceramics Piezopolymers Piezocomposites
k
t
(%) 45-55 20-30 60-75
Z (MRayls) 20-30 1.5- 4 4-20
33
0
T


200-5000 ~10 50-2500
Tanδ (%) <1 1.5-5 <1
Q
m
10-1000 5-10 2-50
ρ (10
3
kg/m
3
) 5.5-8 1-2 2-5

28

2.2 Piezoelectric Composites
2.2.1 High Frequency Composites for Linear Arrays
There are many piezoelectric materials available to choose from when designing a single
element or array transducers. Among those currently available, piezocomposites, which consist
of an array of piezoelectric pillars embedded in a passive polymer matrix, are widely used in
fabricating ultrasound imaging devices due to their enhanced electromechanical coupling
efficients (k
t
) and acoustic impedances, which are better than those of solid piezoelectric
ceramics and crystals (Smith, 1992). Another advantage of piezocomposites is unlike bulk
piezoelectrics, their material properties, such as electrical and acoustic impedance, can be
tailored to specific requirements. Finally, the need for an acoustic lens can be eliminated as
piezocomposites are typically more flexible and can easily be shaped into forms. This is
advantageous because commonly used lens materials such as urethane and epoxy can be very
attenuative at high frequencies, and for array transducers, the lens can also cause significant
degradation during beamforming.
Although the piezocomposites have several advantages over bulk piezoelectrics, their
benefits come at the expense of introducing more undesired inter-pillar resonances. Auld and
Smith demonstrated that the resonances may be due to the formation of Lamb waves in the
piezocomposites’ periodic microstructure. These resonances couple strongly with the
fundamental thickness-mode resonance frequency and hence degrade the performance of the
device; this is especially true when using the composites to fabricate linear array transducers
since they are more susceptible to cross-talk (mechanical and electrical coupling) between array
elements. The resonances cause elements not to operate independently but together, leading to

29

changes in the beam pattern, electrical impedance, and echo response (R. L. Baer, 1984, D.
Ccrton,2001, and C. E. Démoré, 2006).
The coupling coefficient of a piezoelectric material is dependent on the geometry of the
material. A thin disk-shaped single element transducer with a coupling coefficient, k
t
, can be
improved by replacing its monolithic ceramic element with a series of either long thin elements
with a coupling coefficient, k
33’
, or tall narrow elements (k
33
) imbedded in a polymer matrix. The
resulting piezoelectric composite element will display a higher effective thickness mode
coupling coefficient, reduced lateral mode coupling across the width, and lower acoustic
impedance. In the late of 1970’s, a variety of piezocomposite structures, including 0-3, 1-3, 2-2,
3-1, 3-2, and 3-3 connectivity, was first introduced and examined by Newnham. He described the
importance of connectivity on piezocomposite performance. Connectivity greatly impacts the
performance of a composite because it determines if either the piezoceramic or the polymer
dominates the elastic, dielectric, and piezoelectric properties of a given geometry. The
connectivity of a composite is defined as the number of dimensions in which each component
phase is continuous. Among these, 1-3 connectivity piezocomposites have become the most
common because they are well suited to underwater, biomedical and nondestructive testing
applications. The PZT-5H 1-3 piezocomposites, for example, have a higher electromechanical
coupling coefficient (K
33
=0.75) than that of both the PZT-5H 2-2 piezocomposites (K
33’
=0.65)
and the traditional PZT-5H ceramic disk (K
t
=0.5). (see Figure 2.1) The 1-3 composites consist
of posts of piezoceramic embedded in a polymer matrix, and the 2-2 composites consist of side-
by-side strips of piezoceramic and polymer. (see Figure 2.2) The 1-3 refers to that the piezo-
ceramic is continuous in only the z-direction whereas the polymer is continuous in the x, y and z-
directions, hence the designation of “1-3” composite. The standard notation that describes a

30

composite is therefore “X-Y” where X is the piezoceramic connectivity and Y is the polymer
connectivity.

Figure 2.1 Piezoelectric geometries and associated electromechanical coupling coefficient for
single element transducers, 2-2 composites and 1-3 composites transducers.

31

Figure 2.2 Two different configurations of piezoceramic composites: (a) 1–3 composite; (b) 2–2
composite.

When designing or using composites the spatial scale of the materials used must be
considered. Only the thickness mode of the piezo-ceramic must be able to be excited within the
transducer passband range in order to produce a broadband, short, impulse response. If other
resonance modes are excited within the composite, the thickness mode coupling of the device
will degrade, and undesired resonances will appear in the time domain response. These

32

resonances can be considered the product of lateral resonances caused by the widths of the piezo-
ceramic and polymer elements. A number of researchers have investigated these lateral
resonances as they relate to composite performance (Smith and Auld, 1991; Oakley, 1991;
Certon et al., 1997).
Reynolds et al. first investigated the generation of these resonances within 1-3
piezocomposites and showed the first two “Lamb” mode frequencies, f
L1
and f
L2
, can be
predicted fairly well using the equations given
1 L
s
C
f
d
 (2.8)
2
2
Ls
C f
d
 (2.9)
where C
s
corresponds to the shear sound velocity in the polymer and d is the pillar-to-pillar
spacing within the composite. The driving force behind these inter-pillar resonances is the
piezoelectric composite’s thickness mode displacement. Therefore, these resonances will always
exist but will only become prominent when their frequencies approach those of the fundamental
thickness mode and its harmonics. Many models have been used in an attempt to define the
maximum allowable spatial scale for periodic composites including finite element modeling
(Hayward and Bennet, 1996; Qi, 1997) and dynamic modeling of guided waves (Geng, 1997). In
the absence of these models a number of rules can be used to approximate the largest allowable
spatial scale of PZT-polymer composites. These rules, proposed by Ritter (2000), are written
using the criteria that all lateral resonances must be kept above twice the device center frequency
(Smith, 1989; Oakley, 1991; Geng 1997).

33

For the 2-2 composites with low to intermediate piezoceramic volume fractions, the first
piezoelectrically coupled lateral resonance is determined by the half wavelength resonance for a
shear wave that fits in the kerf (Oakley, 1991; Geng, 1997). For 2-2’s whose has a high
piezoceramic volume fraction, the first lateral resonance occurs as a result of a half wavelength
resonance for a longitudinal wave across the width of the ceramic (Oakley, 1991; Geng, 1997).
Therefore in order to push both of these resonances sufficiently above the device passband the
following criteria must be met:
_
4
s
c
V
Kerf width
F


(2.10)
_
4
l
c
V
Creamic width
F


(2.11)
where Vs is the shear wave velocity in the polymer kerf, and F
c
is the device center frequency,
and V
l
is the longitudinal wave velocity for the width resonance of the ceramic.
For a 1-3 composite, the first lateral mode resonance is usually caused by a shear wave
through a polymer that fits diagonally between two ceramic pillars in the kerf. In order to avoid
this resonance the following criteria should be considered (Ritter 2000):
_
42
s
c
V
Kerf width
F


(2.12)
where Vs is the shear wave velocity in the polymer kerf, and F
c
is the device center frequency.
The piezoceramic can also contribute to the lateral resonance, and therefore the piezoceramic
height should be much larger than the width. A general rule of thumb is that a width-to-height

34

ratio of less than 0.5 is sufficient to avoid this mode. This is, of course, dependent on the
piezoelectric material used.
As the operating frequency of a transducer goes up (> 20MHz), the lateral interference
and spurious modes, which are caused by lateral resonances and the pillar-to-pillar periodicities,
respectively, more likely lead to image artifacts, such as “ghost” images (i.e., attenuated and
slightly delayed replica overlaid on top of the real image). The conventional solution to this
problem is to make the kerf- and pillar-widths sufficiently small so as to ensure that the
frequencies of these lateral resonances are at least double that of the transducer's fundamental
frequency. However, this solution cannot be applied to the 1-3 piezocomposites with a 30 MHz
operating frequency or higher because of that the required kerf is less than 10 μm. As a result,
fabricating a 1-3 piezocomposite material for high frequency applications is very challenging.

2.2.2 Fabrication of High Frequency 1-3 Composites
In 1981, H.P. Savakus first manufactured and documented the 1-3 piezocomposites using
the traditional “dice-and-fill” technique, in which a plate of piezoelectric material is diced using
a mechanical saw, and a polymer is applied and then cured within the kerfs. This method has
become the most widespread and standard method for fabrication of 1-3 piezoelectric composites
because of its simplicity and high efficiency. In the past two decades, the dice-and-fill methods
have been commercially used in building piezocomposite transducer at the clinical frequencies in
the lower 1-10 MHz range. However, as the frequency of operation increases, the desired kerf
and pillar-widths decrease sharply; for instance, the required kerf for a 20 MHz 1-3
piezocomposite is less than 10 μm, which is smaller than the width of the thinnest blade currently

35

available for commercial dicing saws. Therefore, for the mechanical dice-and-fill technique, the
fabrication of fine-scale composites which are immune to the deleterious effects of lateral
resonances is still very challenging.
An essential parameter for designing the 1-3 piezocomposites is the relationship of the
lateral spatial scale (the pitch) and the plate’s thickness (height). One proposed criterion (P.
Reynolds, 2003) states that the pillar width-to-height ratio must be less than 0.5 to avoid
interference from the lateral resonances (inside the pillar) and spurious modes. Ideally, for a 40
MHz PZT-5H 1-3 piezocomposite transducer, it requires a kerf width of approximately 6 μm is
to push the first lateral frequency to about 80 MHz (twice the fundamental frequency), while
maintaining a width-to-height ratio inferior to 0.5 and a ceramic volume fraction superior to 50%.
However, the smallest blade width on the commercial dicing saw is around 15μm. Thus, the
traditional dice-and-fill technique needs to be improved or even replaced in order to work for the
high-frequency 1-3 piezocomposites.
In addition, since the piezo-composite properties are largely determined by the shear
properties of the filler epoxy, Brown et al. used a very soft epoxy as the matrix to significantly
minimize the lateral resonances. This soft epoxy provided good pillar-to-pillar damping across
the composite; however, the soft epoxy was not able to resist high voltage during poling, causing
the device to short-circuit. This drawback is especially critical for the high-frequency composite
array transducers, which typically have numerous fine-scale kerfs.
Many alternative methods have been employed to produce such an ultrafine scale
piezocomposite. Generally, they can be divided into two major categories. First is the removal of
piezoelectric material, including laser dicing (M. Lukacs, 2006) and dry etching (J. R.

36

Yuan,2006). M. Lukacs used laser instead of a mechanical dicing saw to make a 40 MHz
ultrasound transducer with a 33 μm pitch and an 8 μm kerf, whilst laser machining requires
prohibitively long machining times and suffers from cross sectional tapering and material re-
deposition. The dry etching method has been used by J. R. Yuan at al. to fabricate a 40 MHz 1-3
composite transducer with a 17 μm pitch and a kerf of 4 μm. However, the drawbacks of this
method are slower etching rates and higher maintenance fees.
The second category is the formation of piezoelectric material, such as micro-moulding
(S. Wang, 1999), tape casting (W. Hackenberger, 2000), and micro-printing (A. R. Bhatti,2001).
The micro-moulding method may provide the PZT composite with the thickness from 12 to 100
μm and a high aspect ratio. A 26 MHz 1-3 PZT piezocomposite transducer was reported by S.
Cochran et al. in 2004. Recently, a new micro-molding technique for fabricating high-frequency
transducers has been reported. The resulting structures can be fabricated with aspect ratios up to
3:1 and thicknesses up to 50 μm with a 0.3 k
t
. The tape casting method can be used to fabricate
the 2-2 and 1-3 piezocomposites. S. Kwon has reported a 20 MHz 2-2 piezocomposite transducer
with a pillar width of 25 μm and an epoxy-filled kerf width of 5 μm. Using the micro-printing
method, R. Noguera in 2005 fabricated a 1-3 piezocomposite with 50% ceramic volume ratio and
90 μm pillar width.
However, various issues have prevented commercial adoption; for instance, injection
moulding incurs high tooling costs and is susceptible to dimensional changes during the
demoulding of fine structures. The tape casting method is good at making the 2-2
piezocomposites but not the 1-3 piezocomposites. The micro-printing method may not be able to
be used to make a 1-3 piezocomposite with a center frequency of 20 MHz or higher.

37

Consequently, these processes impose constraints on volume fraction, minimum dimension,
materials quality, cost, time, or repeatability of manufacture, thus limiting their commercial
usefulness.
In order to develop alternate methods of lateral resonance suppression, from 1991 to 1996,
Hossack and Hayward et al. compared low-frequency piezocomposites of different pillar
geometries. They found that triangular pillars have the potential for improved electromechanical
characteristics when the pillars are oriented to avoid facing parallel surfaces. Furthermore, for
the 1-3 triangular pillar composite, when its surface dilation homogeneity (a measure of how
uniformly the pillar-epoxy surface vibrates in the thickness mode) decreases below 90%, the
efficiency of the composite decreases severely. Therefore, to keep the surface dilation
homogeneity above 90%, Hayward proposed a set of criteria, such as a maximum pillar aspect
ratio (MPAR), to push the frequency of the first lateral resonance to at least 2 times the center
frequency of the fundamental mode. Generally, an MPAR of approximately 0.4 was then
required for a triangular pillar composite with a 40% ceramic volume fraction (G. Hayward,
1996). However, this rule is not suitable for high-frequency mechanically diced 1-3 composites
since it makes the fabrication of the 1-3 composites extremely difficult. In 2007, Brown et al.
developed a 1-3 piezocomposite with mechanically diced triangular pillars (MPAR > 0.8) to
reduce the influence of the lateral-mode resonances on the thickness-mode resonance within the
composite by spreading the energy over a broad frequency range. Compared with the square
pillar composites, the triangular pillar composites showed a 9.5 dB reduction in spurious signals
and a 30% gain in the 2-way pulse bandwidth. In 2010, Yin et al observed that lateral resonance
suppression in piezocomposites with 45° triangular pillars was better than that of 30° and 60°

38

triangular pillars. He attributed this result to a higher level of geometric complexity of 45°
triangular pillars.
The above observations imply that the highest level of geometric complexity translates to
the greatest suppression of coupling between lateral- and thickness-mode resonances.
Theoretically, the highest level of geometric complexity can be obtained using piezocomposites
with distributed period pitches, such as randomly spaced/sized elements (J. A. Hossack, 1992, J.
Yuan, 1996, and G. Harve, 2006). Hossack and Hayward tested this idea on several low-
frequency 1-3 piezocomposites with a distributed period structure and demonstrated no strongly
coupled resonant frequencies. The aforementioned studies inspired us to develop and test a
composite with a pseudo-random pillar geometry and distribution that could be used in the
design of ultrasound transducers operating at frequencies up to 40 MHz. In addition, this
composite geometry was developed so that a mechanical dicing saw could still be used for
fabrication.
The following chapter describes the design and fabrication of the 1-3 piezocomposites
with pseudo-random pillars. For comparison, using a modified mechanical dicing technique, we
designed and fabricated several PZT-5H based 1-3 composites with different pillar geometries
(square (SQ), 45° triangle (TR), and pseudo-random (PR). Using these composites, several
ultrasound transducers were fabricated for the comparison of their acoustic performance.

39

Chapter 3: Composite Fabrication and Evaluation
This section outlines the procedure used to fabricate 1-3 composites with pseudo-random
pillars. The impact of element aspect ratio on the coupling coefficient and elastic stiffness is
discussed. A finite element modeling is then performed to predict the performance of the pseudo-
random composites. Measured acoustic properties of the composites are presented and
summarized.

3. 1 Composites Design
The different pillar geometries (square, 45° triangle, and pseudo-random) fabricated for
this study were designed to have a ceramic volume fraction of nearly 50% and kerf width, K, of
14µm (see Table 3.1). This design enabled us to compare the performance of these composites
fairly since the ceramic volume fraction of a composite determines the density and the
permittivity. In order to keep the same ceramic volume fraction and kerf width at the same time
for all the pillar geometries, the pillar widths, W, of the SQ, TR and PR composites had to be
kept at 34 µm, 38 µm, and 57 µm, respectively. The thickness, H, which corresponds to the
antiresonant frequency of 10 MHz was 165 µm and at 15 MHz was 110 µm.
When considering materials for array elements the transducer engineer must look
carefully at the impact of the element aspect ratio on both the coupling coefficient and elastic
stiffness. Coupling between the width and height resonances for an element can degrade the
response of the element. With a poor aspect ratio, excitation of the height extensional mode will
result in a loss in sensitivity in the thickness mode. More importantly, since the width or x-

40

direction is usually poorly damped, coupling to it will result in an undamped resonance
appearing in the impulse response of the element. It is therefore important to be able to predict
the allowable aspect ratio for an array element prior to fabrication.

Table 3.1 Designed parameters for the PZT-5H 1-3 composites with different pillar geometries
for this study.

C.F.
(MHz)
Thickness(H)
(µm)
Max.
Aspect Ratio
(W/H)
Pillar
Width
(W)
(µm)
Kerf (K)
(µm)
Pitch (P)
(µm)
Ceramic
V olume
Fraction
(%)
Square 10 165 0.29 34 14 48 50
45°
Triangle
10 165 0.33 38 14 52 46
Pseudo
Random
10 165 0.35 57 14 71 49
Square 15 110 0.44 34 14 48 50
45°
Triangle
15 110 0.49 38 14 52 46
Pseudo
Random
15 110 0.52 57 14 71 49

One method for looking at the impact of aspect ratios is the mode coupling theory
presented by Onoe and Tiersten, 1963, for the short-circuit case, and DeSilets, 1978, for the
open-circuit case. This theory describes the piezoelectric element as a vibration problem
consisting of two degrees of freedom coupled through a single mechanism.

41

Ideally, for low-volume fraction composites, the pillar width-to-height ratios (W/H) must
be less than approximately 0.5 in order to avoid the deleterious effects of mode-coupling (G.
Hayward, 1996). The higher aspect ratio would lead to significant coupling between the lateral-
and thickness-mode resonances. The maximum aspect ratios (the diagonal direction) of these
composites, therefore, all remained below 0.5 except the PR 15 MHz (0.52).

42

3. 1.1 Finite Element Modeling
Finite element modeling (FEM) is the most common method employed to analyze
complex vibration modes in a structure. In this method, the vibrating continuum is separated into
small elements. A set of discrete relationships can be written and analyzed for the interactions
between these finite elements. Lerch (1990) and Qi (1997) published more detailed description
of finite element modeling of single element and array transducers. PZFLEX (Weidlinger
Associates, Los Altos, CA) is a time-domain modeling software package for analysis of
piezoelectric resonators. The finite element method used in PZFLEX reduces the
electromechanical partial differential equation for the model space to a system of ordinary
differential equations with respect to time. (Wojcik et al., 1993). Hence, the PZFLEX was used
for this study because it was suggested to provide a better means for modeling the time
dependent array element echo response.
3-Dimensional models of the 1-3 composites with different pillar geometries (SQ, TR,
PR) were analyzed to simulate their electrical impedances and phase angles (see Figure 3.1). For
broadband high frequency designs using passive materials with low shear velocities, a small
mesh size (1/15 of a wavelength in the material with the lowest velocity) is required to improve
the modeling accuracy. Figure 3.2 shows the modeled electrical impedance magnitude and
phase angle for the three different pillar geometries at 15 MHz. Compared with the SQ and TR
composites, at the same frequency the PR composites had the least lateral modes. This
encouraging result confirmed our hypothesis about the PR composites before their fabrication.

43

Figure 3.1 3D geometry of the composites with different pillars constructed by PZFLEX.

44

Figure 3.2 Modeled electric impedance and phase angle of the composites with different pillars
using PZFLEX 3D model.

45

3. 2 Composites Fabrication
Commercial PZT-5H ceramic plates (3203HD, CTS Electronic components, Inc.,
Albuquerque, NM) were diced into smaller square pieces (15 mm x 15 mm, 550 µm thick) and
then individually bonded onto flat glass carriers using low-temperature para ffin wax. A
programmable dicing saw (Tcar 864-1, Thermocarbon, Inc., Casselberry, FL) was used to cut
kerfs into these ceramic plates using a 12 µm blade. Since the vibration of the blade cannot be
avoided, the resulting kerf widths in the composites will be slightly larger than the dicing saw
blade thickness. Depending on the state of the blade, the resulting kerf width typically fluctuated
between 12 and 14 µm.
Since the pillar scale was quite close to the machinability limit of the PZT-5H on our
dicing saw, to alleviate the pillar breakage, a double-index-dicing technique was used to produce
such fine-scale pillar spacing (J. M. Cannata, 2006). To keep the volume fraction constant for all
different pillar geometries, different pillar widths and pitches were used; thus, we detail the
fabrication process of each composite pillar geometry separately.
1. Square (SQ)
First, using the 12 µm dicing blade, two sets of cross cuts (200 µm deep) were first made
into the ceramic at right angles to each other with a pitch of 96 µm. These first cuts were
subsequently filled with Epotek 301 epoxy (Epoxy Technology Inc., Bellerica, MA) as a result of
capillary action. The filler in this first set was left to cure at room temperature in a dry nitrogen
environment for three days. The excess epoxy was then lapped o ff to expose the diced ceramic.
The second cuts were then made between these first cuts to produce the final 48 µm composite
pitch. The second set of diced kerfs was subsequently filled in the same manner. After curing and

46

lapping off the excess epoxy from the top of the composite, this composite was flipped over and
then lapped down to 165 µm. The average width of the resultant kerfs was 14 µm, and the width
of the ceramic pillars was 34 µm. The net piezoceramic volume fraction of this square pillar 1-3
composite was 50 % (see Fig. 3.3(a)).
2. 45° Triangle (TR)
As with the square pillar composite, on the top of the ceramic, two sets of perpendicular
cuts were first made in the ceramic 200 µm deep at a 73 µm pitch and then filled with epoxy.
After curing, the second set of cuts was made at 45 degree angles relative to the vertical with a
double modified pitch (2*51 µm) and then filled with epoxy again. After the epoxy had cured,
the third set of cuts was then made between these second set of cuts to produce the final modified
pitch (51 µm). This modified pitch kept the pitch (73 µm) between the first cuts constant over the
entire surface of the composite. The resulting diced kerfs were subsequently filled with epoxy in
the same manner. Since one more cut was required to fabricate the TR composites, it was more
time-consuming than that of the SQ composites. The average resultant kerf width was 14 µm,
and the width of the ceramic pillars was 46 µm with a net piezoceramic volume fraction of 46 %,
which is shown in Fig. 3.3(b). For more detail about the fabrication of a 45° Triangle 1-3
piezocomposite see Yin et al..
3. Pseudo-Random (PR)
The pseudo-random composite fabrication procedure was similar to the aforementioned
processes for the square and the 45° triangle pillars. The differences were that two sets of cross
cuts were first made at angles of 25° and 145°

relative to the horizontal, with a pitch of 71 µm.
This first set of cuts was then filled with Epotek 301 epoxy and cured at room temperature. The

47

second set of cuts was subsequently made at an angle of 90° relative to the horizontal on the
same surface with the same pitch. Since no third cut was required for the PR composites, it was
the least time-consuming of the three composites fabricated. The maximum ceramic pillar width
was 57 µm, and kerf width was also 14 µm. The estimated piezoceramic volume fraction of this
1-3 composite was 49 % (see Fig. 3.3(c)).

Figure 3.3 Different pillar geometries for the PZT 5H 1-3 composites. (a) Square (b) 45°
Triangle (c) Pseudo-Random

48

3. 3 Composites Evaluation
To determine how performance changes when the maximum aspect ratio for each
composite increases, these composites were each cut in half. One half was used to fabricate the
10 MHz transducers, and the other half was then lapped down for the 15 MHz transducers. The
top and bottom surfaces of all the composites were then cleaned and sputtered (NSC-3000,
Nano-Master Inc., Austin, TX) with a total of 1000 Å Cr/Au as a conduction layer. Each sample
was poled in air at room temperature for 10 minutes using a DC field of approximately 3 kV/mm.
The electrical impedance of freestanding composites were measured with an impedance analyzer
(Model # 4294A with 16034H test fixture, Agilent Technologies, Santa Clara, CA); and the
results are shown in Figure 3.5. The thickness mode electromechanical coupling coefficient (k
t
),
clamped dielectric loss tangent tan(δ
s
e
), mechanical quality factor (Q
m
), attenuation coefficient
(α), relative clamped permittivity (ε
S
33
/ε
o
), and longitudinal velocity (V
l
) were calculated using
measurements of air-resonating samples (F. S. Foster,1991 and H. Wang, 2001) and a sub-
micrometer thickness gauge (Heidenhain CT25/ND281B, Schaumburg, IL).
The measured electrical impedances, in Figure 3.4, of the 10 MHz and 15 MHz 1-3
composites were compared. For the SQ and TR 10MHz composites, a minor lateral resonance
was visible at 20 MHz; however, no obvious lateral resonance was found on the PR 10MHz
composites. This might be attributed to the non-periodic structures within the PR composites.
When the frequency was increased to 15 MHz, the fundamental frequency of the SQ composites
combined with the first lateral resonance peak and the rest of lateral resonance peaks were found
from 20 to 30 MHz though for the TR and PR 15MHz composites, fewer lateral resonance peaks
were found in the same frequency range. A comparison of the amplitude of the fundamental

49

frequency peak with that of the highest lateral resonance peak shows that, at 15 MHz, the PR
composites were less affected by the lateral resonance than the SQ and TR composites.

Figure 3.4 Measured electrical impedance (solid line) and phase angle (dashed line) for the
piezocomposites with different pillar geometries. At 10 MHz, square (a), triangle (b),
and pseudo-random (c). At 15 MHz. square (d), triangle (e), and pseudo-random (f).

0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
Magnitude (K )
Magnitude (K )
Square
-90
0
90
0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
-90
0
90
(f) (e) (d)
(c) (b)
Triangle
(a)
0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
Pseudo-Random
Magnitude
Phase Angle
-90
0
90
Phase Angle (degrees)
0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
Frequency (MHz)
-90
0
90
0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
Frequency (MHz)
Frequency (MHz)
-90
0
90

0 10 20 30 40
0.0
0.5
1.0
1.5
2.0
-90
0
90
Phase Angle (degrees)

50

The piezoelectric properties of these composites were calculated from the measured
electrical impedance (see Table 3.2). The average of these properties was obtained from three
samples of each the composite pillar geometries. The results demonstrate that the PR composites
had slightly higher relative clamped permittivity (493 @ 10 MHz and 520 @ 15 MHz). This is
advantageous for the composite transducer with a small aperture size since higher relative
clamped permittivity may better match the 50 ohm electrical impedance of commercial systems.
On the other hand, the PR composites had higher k
t
values (0.62 @ 10 MHz and 0.61 @ 15 MHz)
than the SQ and TR composites. Study provides evidences that, for a free standing PZT bar, k
t

increases to its maximum at around an aspect ratio of 0.6 (Yin, 2005). The fact that PR
composites have greater maximum aspect ratios (0.35 at 10 MHz and 0.52 at 15 MHz) may lead
to the higher k
t
values. Moreover this enhanced k
t
value may be attributed to that there is more of
a distribution of pillar resonance frequencies due to the varying pillar widths.
Additionally, although the maximum aspect ratio for the PR composites was higher than
those of the SQ and TR composites, the PR composites still out performed them. This
observation implies that the PR composites may have a higher limit for the maximum aspect
ratio. This would be advantageous for fabricating high-frequency transducers. Therefore, the PR
composites may not only suppress the lateral interference better than the SQ composites, but also
provide a higher k
t
than the TR composites. To further investigate their frequency response
properties, we measured pulse-echo results of single element transducers fabricated from these
composites.

51

Table 3.2 Average measured material properties for the PZT-5H 1-3 composites with different
pillar geometries manufactured for this study.

Thickness
(µm)
k
t
tan
δ
Q
m

α
(dB/mm)
ε
S
33
/ε
o

V
l
(m/sec)
SQ
10M
164
(0.5)
0.57
(0.012)
0.12
(0.009)
44.8
(3.8)
1.85
(0.15)
472
(25)
4037
(35)

TR
10M
165
(0.7)
0.52
(0.018)
0.20
(0.013)
41.9
(3.1)
1.97
(0.28)
484
(33)
3653
(42)

PR
10M
161
(0.3)
0.62
(0.020)
0.15
(0.008)
33.1
(4.3)
2.84
(0.22)
493
(15)
3850
(39)

SQ
15M
111
(0.7)
0.58
(0.011)
0.11
(0.010)
21.6
(2.9)
5.68
(0.19)
501
(37)
4173
(55)

TR
15M
113
(0.8)
0.51
(0.023)
0.20
(0.007)
22.7
(3.7)
5.31
(0.25)
457
(29)
3526
(47)

PR
15M
110
(0.5)
0.61
(0.019)
0.16
(0.015)
25.1
(3.5)
4.93
(0.28)
520
(45)
4015
(59)

Values in parentheses are the standard deviations for the measured properties.

52

Chapter 4: Composite Transducer Fabrication and evaluation
This chapter discusses the fabrication process of composite transducers. Several single
element transducers were made to investigate the performance of 1-3 composites. An electrical
impedance analyzer was used to measure the electrical impedance of the composites. The center
frequency, bandwidth, pulse length and relative sensitivity for the composite geometries were
each tested using a pulse-echo setup. Insertion loss of the composite transducers were measured
and compensated for losses due to diffraction and attenuation in the water bath.

4.1 Single Element Transducer
4.1.1 Single Element Transducer Fabrication
Single element transducers were fabricated first by bonding conductive epoxy (E-
SOLDER 3022, Von Roll Isola Inc., CT) backing material to the composite. The resulting plates
were mechanically shaped into 2.5 mm disks using a lathe and then made into single element
ultrasound transducers. A prefabricated Epotek 301 epoxy sheet (Z=3.04 Mrayl, 66 μm for 10
MHz and 44 μm for 15MHz) matching layer was bonded to the front face of the composite
transducers. A cross-sectional drawing of the single element ultrasound transducer is shown in
Figure 4.1 More details about the design and fabrication of this ultrasound transducer were
reported elsewhere (J. M. Cannata, 2003).

53

Figure 4.1 A cross-sectional drawing of the single element ultrasound transducers.
4.1.2 Single Element Transducer Evaluation
The center frequency, bandwidth, pulse length, and two-way sensitivity of each
composite transducer were measured by a common two-way pulse-echo test. A Panametrics
5900 pulser/receiver (Olympus NDT Inc., Waltham, MA) was used to excite each transducer at
the 1 µJ, 50 Ω settings with no gain. The transducer was positioned in a degassed/deionized
water bath opposite a flat quartz reflector. The echoes from the transducer were then digitized
and displayed with a 500MHz oscilloscope (LC534, LeCroy Corp., Chestnut Ridge, NY). (see
Figure 4.2) Fast Fourier transform (FFT) of the echo yielded the frequency response of the echo.
The upper and lower frequencies of the signal corresponded to the first and the second −6 dB
points of this power spectrum, respectively. The mean of these frequencies was recorded as the
SMA Connector
Brass Housing
Conductive Backing
Insulating Epoxy
Piezo-composite
Epoxy Matching Layer
Cr/Au Electrodes

54

center frequency of the transducers, and dividing the di fference of the frequencies by the center
frequency gives the bandwidth of the transducers. The amplitude of the echo signal was recorded
for relative element sensitivity comparisons. The length of time between the first and last points
where the signal was -20 dB relative to the peak was recorded as the -20 dB pulse length of the
echo waveform.
The measured pulse-echo waveforms, their spectrums and envelopes (see Figure 4.3) of
these composite transducers showed that, for instance, the SQ 10MHz device had the longest -
20 dB pulse length (282 ns), and the TR 10MHz one had a shorter -20 dB pulse length (239 ns)
but lower sensitivity (0.48 V). Note that the PR 10MHz device had the shortest -20 dB pulse
length (230 ns) and the strongest sensitivity (0.60 V). At 15 MHz, the -20 dB pulse length of the
TR 15MHz device was 140 ns, which was about 30 and 10 ns less than that of the SQ 15MHz
and the PR 15MHz, respectively, but the PR 15MHz still had the strongest sensitivity (0.75 V).
This observation indicates that the TR and PR composites both were able to reduce the -20 dB
pulse length better than the SQ composites; and the PR composites were the most sensitive.
Besides, no matter with or without a matching layer, the TR and PR composites have wider -6
dB bandwidth that the SQ composites. Table 4.1 also summarized the measured frequency
response properties for theses composites transducers with different pillar geometries.

55

Figure 4.2 Block diagram of a pulse-echo measurement, which is the most common test for
transducers.

56

Figure 4.3 Measured pulse-echo waveform (a), their spectrums (b), and envelopes (c) at 10 MHz
for the composite single element ultrasound transducers. Measured pulse-echo
waveform (d), their spectrums (e), and envelopes (f) at 15 MHz for the composite
single element ultrasound transducers.
14.6 14.8 15.0
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
21.6 21.8 22.0
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
0.0 5.0 10.0 15.0 20.0 25.0 30.0
-30
-24
-18
-12
-6
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0
-30
-24
-18
-12
-6
0
14.6 14.8 15.0
0.0
0.2
0.4
0.6
0.8
1.0
21.6 21.8 22.0
0.0
0.2
0.4
0.6
0.8
1.0
Time( s)

SQ10MHz
TR10MHz
PR10MHz
Time( s)
SQ15MHz
TR15MHz
PR15MHz

Amplitude(mV)
(e)
(d)
(b)
Amplitude(dB)

Frequency(MHz)
(a)

Frequency(MHz)
-20 dB
(f)
Time( s)
Amplitude(normalized)

Time( s)
-20 dB
(c)

57

Table 4.1 Measured frequency response properties for the PZT-5H 1-3 composites transducers
with different pillar geometries.

Center
Frequency
(MHz)
Without a
macthing
layer, -6 dB
bandwith
(% C.F.)
With a
macthing
layer, -6 dB
bandwith
(% C.F.)
V
pp
(V)
-20 dB
pulse length
(ns)
SQ 10M 10.5 48 54 0.48 282
TR 10M
11.6 67 65 0.41 239
PR 10M
11.8 60 61 0.60 230
SQ 15M
15.8 55 58 0.62 172
TR 15M
16.0 65 69 0.48 140
PR 15M 15.9 63 66 0.75 148

58

4.2 Array Transducer Sub-aperture
4.2.1 Array Transducer Sub-aperture Fabrication
Next, we fabricated 15 MHz sub-aperture arrays consisting of 16 elements and
approximately 0.95 λ pitch (144 µm) from these composites to measure the acoustic cross-talk
between elements. The orientation of elements relative to composite for each type is shown in
Figure 4.4. We separated the array elements by mechanically dicing off the top electrode layer
over the composite kerfs to the depth of 1 µm with a 10 µm wide dicing blade. Subsequently, we
used an intermediate flexible circuit incorporating 5-µm-thick copper traces on a 25-µm-thick
polyimide surface to connect individual coaxial cables to each array element. The flexible circuit
was connected to the composites each by carefully aligning and bonding the copper traces to the
top electrodes with Epotek 301 epoxy. This interconnected array assembly was allowed to cure
for 48 hours at room temperature in a dry nitrogen environment before further processing.
After curing, an additional 1000-Å Cr/Au electrode was sputtered to connect the ground
side of the composites to the flexible circuit. A 44 µm thick prefabricated Epotek 301 epoxy
sheet was bonded to the top surface of the composites as a matching layer. Another prefabricated
backing block made of the same epoxy (7 mm thickness) was then bonded to the other side of the
flexible circuit. The 7 mm backing thickness used was deemed acceptable to avoid backing echo
interference based upon previously reported acoustic attenuation data (H. Wang, 2001).
Therefore, from top to bottom, the four layers of these sub-aperture arrays are the epoxy
matching layer, the composite piezo-layer, the flexible circuit, and the epoxy backing layer.
Then, 1-m long, 75-Ω coaxial cables were soldered to the flexible circuit with a low-temperature

59

indium-based solder (Indium Corporation of America, Utica, NY). A picture of the completed
array is shown in Fig. 4.5.

Figure 4.4 Sputtered 1-3 composites with different pillar geometries (removing the Cr/Au over
the epoxy kerfs) (a) Square (b) 45° Triangle (c) Pseudo-Random.

Figure 4.5 A picture of the completed array sub-aperture.

60

4.2.2 Array Transducer Sub-aperture Evaluation
Identical coaxial cables were connected the array elements to the electronics to provide
equal loading conditions. These array sub-apertures were then placed in a deionized water bath
with a flat quartz reflector. Since no dead element was found in these array sub-apertures, eight
elements (Element 1 to 8) of each the array sub-apertures were selected to represent all the
elements.

4.2.2.1 Pulse-echo measurement
An identical two-way pulse-echo test was performed to measure pulse length of each
array transducer sub-aperture. The Panametrics 5900 pulser/receiver (1 µJ, 50 Ω settings with no
gain) was used to excite each transducer. The same flat quartz plate in a degassed/deionized
water bath was used as a reflector. The echoes recorded by the transducer were then digitized and
displayed with the LC534 oscilloscope. (see Figure 4.2) The frequency response of the echo was
converted from the measured echo signal using Fast Fourier transform (FFT). The pulse length
of the echo waveform was recorded as the length of time between the first and last points where
the signal was – 20 dB relative to the peak. Table 4.2 shows the measured - 20 dB pulse length
for the 15 MHz array sub-apertures. The PR array sub-aperture had the shortest -20 dB pulse
length of 143.8 ns (SQ: 202.7 ns; TR: 151.2 ns).

61

Table 4.2 Measured insertion lose (IL) and – 20 dB pulse length (PL) for the 15 MHz array sub-
apertures with different pillar geometries. Element 1 to 8 of each the array sub-
apertures were measured to represent all the elements.

1 2 3 4 5 6 7 8 Mean

SQ

IL (dB@15 MHz)
-25.2 -26.1 -26.6 -24.8 -25.5 -24.8 -27.0 -25.5 -25.6
(0.81)

-20 dB PL (ns)
195 206 205 193 210 202 198 213 202.7
(7.08)

TR
IL (dB@15 MHz)
-28.3 -27.5 -26.9 -27.2 -26.6 -27.7 -29 -28.1 -27.6
(0.78)

-20 dB PL (ns)
156 145 140 153 153 142 163 158 151.2
(8.13)

PR
IL (dB@15 MHz)
-24.5 -25.3 -24.2 -24.8 -25.3 -22.3 -24.5 -22.8 -24.2
(1.10)

-20 dB PL (ns)
142 150 148 155 144 135 128 149 143.8
(8.77)

Values in parentheses are the standard deviations for the measured properties.

62

4.2.2.2 Insertion loss measurement
Insertion loss is the ratio of the output power of the transducer to the input power
delivered to the transducer from the source electronics. The two-way insertion loss for each
element was recorded at 15 MHz. The amplitude of the sinusoidal signal (5 V
pp
amplitude, 20
cycles) from an arbitrary function generator (AFG 3251, Tektronix Inc., Richardson, TX), set in
the burst mode, was measured (without the transducer) across a 50 Ω oscilloscope load at
discrete frequencies from 3 to 30 MHz. These measured voltage served as the reference voltage.
Each array element was then excited by the same function generator; and the echo signal peak
amplitude was recorded at a distance of 11.5 mm (natural focus for the center element) on the
oscilloscope across a 1 MΩ load. (see Fig. 4.6) Measured two-way insertion loss was calculated
using the ratio of the frequency spectrum of the transmitted and received responses, which
compensated for the attenuation in the water bath (2.2 × 10-4 dB/mm.MHz
2
) and loss caused by
the imperfect reflection from the quartz target (1.8 dB). The average compensated insertion loss
for the PR array sub-aperture (-24.2 dB) was lower than that of the SQ and TR array sub-aperture
(-25.6 dB and -27.6 dB, respectively), which were shown in Table 4.2.

63

Figure 4.6 Block diagram of an insertion loss measurement, which is the most common test to
evaluate the sensitivity of a transducer.

64

4.2.2.3 Crosstalk measurement
The same function generator (AFG 3251, Tektronix) was set at sinusoid burst mode (5
V
pp
amplitude) to excite each element of the array. Each element was excited at a step frequency
of 1 MHz throughout the passband (3 to 30 MHz), and the peak applied voltage was recorded as
a reference using the aforementioned oscilloscope set at 1 MΩ coupling. The echo signal peak
amplitudes for all eight elements were measured at a distance of 11.5 mm. Next, after the quartz
reflector was removed, the voltages on the nearest-neighbor and next-nearest elements of each
the element were also measured with the same oscilloscope (also set at 1 MΩ coupling) and
compared with the reference voltage to determine the level of crosstalk [12]. (see Figure 4.7)
For each element, its combined electrical and acoustical crosstalk was measured between its
adjacent element and its next-nearest element. Figure 4.8 shows average measured crosstalk
(electrical and acoustical) between the nearest and next-nearest elements of these 15 MHz array
sub-apertures. The crosstalk was averaged from the measurements of eight elements (Element 1
to 8) of composite pillar geometry. Table 4.3 also summaries average crosstalk at 15 MHz and its
standard deviation of each the composite pillar geometries. For the nearest elements, at 15 MHz
the SQ array sub-aperture, as expected, had the highest level of crosstalk (-25.3 dB). Compared
with the SQ array sub-aperture, the TR array sub-aperture reduced its combined crosstalk by 4.3
dB to -29.6 dB. This observation also validated Brown and Yin’s results [18], [19]. Notably, the
PR array sub-aperture had the lowest measured crosstalk of -31.8 dB, which was 6.5 dB and 2.2
dB lower than that of the SQ and the TR array sub-apertures at the same frequency, respectively.

65

Between the next-nearest elements of these array sub-apertures, a similar reduced-
crosstalk effect was also found. At 15 MHz, for instance, the PR array sub-aperture showed a
combined crosstalk of -43.9 dB, which was 5.5 dB and 2.6 dB lower than that of the SQ and the
TR array sub-apertures, respectively. Therefore, as expected, the PR pillar geometry significantly
reduced the crosstalk between array elements.

Figure 4.7 Block diagram of a cross-talk measurement for array elements.

66

Figure 4.8 Average measured the combined electrical and acoustical crosstalk between the
nearest and next-nearest elements of the array sub-apertures with different pillar
geometries.

Table 4.3 Average measured the combined crosstalk (electrical and acoustical) for the 15 MHz
array sub-apertures with different pillar geometries. Element 1 to 8 of each the array
sub-apertures were measured to represent all the elements.

0 3 6 9 12 15 18 21 24 27 30 33
-60
-55
-50
-45
-40
-35
-30
-25
-20
SQ15
TR15
PR15
Nearest
Next Nearest
Amplitude(dB)

Frequncy(MHz)

Adjacent elements
at 15MHz (dB)
Standard deviation
(dB)
Next-nearest
elements at 15MHz
(dB)
Standard deviation
(dB)
SQ -25.3 0.48 -38.4 0.36
TR -29.6 0.61 -41.3 0.32
PR -31.8 0.65 -43.9 0.43

67

Chapter 5: High-Frequency Kerfless Annular Array Transducer
This chapter describes the design and fabrication of a high-frequency kerfless annular
array transducer utilizing a novel 1-3 piezocomposite which was designed to reduce inter-
element cross-talk. As we know 1-3 piezocomposites have significant advantages over bulk
piezoelectric materials and other types of piezocomposites; however, their benefits come at the
expense of introducing more undesired inter-pillar resonances. At high frequencies, this is
especially detrimental to kerfless annular array transducers. We have previously shown that this
unwanted coupling effect (high inter-element crosstalk), can be further reduced by employing a
pseudo-random pillar geometry. Therefore, we developed 40 MHz annular array transducers
utilizing 1-3 composites with pseudo-random pillars to improve inter-element cross-talk of
kerfless annular arrays.

5.1 Annular Array Transducer
Annular array fabrication is quite challenging because of the dimensional constraints
imposed by the need for operation at high frequencies. Active material selection and preparation
are important to ensure the desired array sensitivity and bandwidth, and inter-element crosstalk
should be minimized so as not to limit the ability of the array to dynamically focus. In a study
conducted by Snook et al., 2004, a 6-element, 45-MHz annular array was fabricated using a fine-
grained lead titanate (PbTiO3). Individual elements were created by laser dicing and connected
by the conventional cable soldering. This work illustrated the advantage of using an active
material with a low planar coupling coefficient to minimize acoustic crosstalk. Brown et al.,
2004, fabricated a 50-MHz, 7-element, kerfless annular array by patterning concentric aluminum

68

electrodes on a planar PZT-5H substrate and using a wire-bonding process for the
interconnections. This work demonstrated that images with excellent lateral resolution, depth of
field, and low side lobe levels could be produced with a kerfless annular array design. Gottlieb
(2005) also fabricated a kerfless 28-MHz composite annular array by bonding a 1–3 composite,
without additional patterned signal electrodes, to a two-sided flex circuit. It was postulated that
the uncharacteristically high average round-trip insertion loss (32.5 dB) observed for this array
was caused by the dominance of the series capacitance generated by the epoxy bond-line on the
electrical impedance of the array elements. In 2011, Hamid fabricated a kerfless 35 MHz 8-
element composite annular array using an interdigital bonded 1–3 composite with square pillars.
The following section describes the design and fabrication of an 6-element kerfless high-
frequency annular array using a similar approach as described Hamid[12]. The major differences
are that a 1–3 composite material with a pseudo-random pillar geometry and the “dice-and-fill”
technique was used. Table 5.1 lists the initial target specifications for this annular array.

69

Table 5.1 Specifications for the 40 MHz 1–3 composite annular array transducer.

5.2 Array Transducer Design and Fabrication
The pseudo-random pillar piezocomposites (33 % ceramic volume fraction, 12 µm kerf
width) were fabricated using a traditional dice-and-fill method. Fine grain PZT-5H ceramic
(TFT-L201F, TFT corp., Tokyo, Japan) was chosen to alleviate dicing damages. This annular
array was designed to have six equal-area elements and a center frequency of 40 MHz. An
intermediate flexible circuit incorporating 5-µm-thick copper traces on a 25-µm-thick polyimide
surface was used to connect individual coaxial cables to each array element (see Figure 5.1).
This flex-circuit was carefully aligned to the composite and bonded using EPO-TEK epoxy 301
in place by applying a constant force to the assembly. The assembly was placed in a dry nitrogen
environment and left to cure overnight. The resulting assembly is shown in Figure 5.2. After
curing, the assembly was cleaned thoroughly and an additional 1000-Å Cr/Au electrode was
sputtered to connect the ground side of the composites to the flexible circuit. A 20 µm thick
prefabricated Epotek 301 epoxy sheet was bonded to the top surface of the composites as a
1–3 pseudo-random pillar composite annular array transducer
Center frequency 40 MHz
Number of elements 6
Aperture size 2.9 mm
Transit focus 4.5 mm
-6 dB Bandwidth 50%
Cross-talk (adjacent elements) < - 30 dB
Insertion loss < 20 dB

70

matching layer. Another prefabricated backing block made of the same epoxy (2 mm thickness)
was then bonded to the other side of the flexible circuit. Identical coaxial cables were connected
the array elements to the electronics to provide equal loading conditions. To provide RF
shielding, the array was encased in a brass tube closed off with a brass end piece. The gap
between the cylindrical brass housing and the transducer assembly was filled by Epotek 301
epoxy. In Figure 5.3, a schematic section drawing of the annular array shows all major
components. Figure 5.4 shows a picture of the completed 40 MHz annular array transducer.
Electrical impedance, pulse-echo, and cross-talk measurements were then employed to evaluate
its performance.

71

Figure 5.1 AutoCAD drawing of 6-element flexible circuit and array elements.

Figure 5.2 The diced PR composites with it flexible circuit. Epoxy 301 was used to glue the
backed composite to the assembly.

72

Figure 5.3: A schematic section drawing of the annular array showing all major components.

Figure 5.4: A picture of the completed 40 MHz annular array transducer.

73

5.3 Array Transducer Evaluation
Imaging is the ultimate experiment to determine the effectiveness of array transducers.
Imaging is affected not only by the array characteristics, but also by the imaging system
electronics and signal processing procedures used. Several standard non-imaging tests were
initially performed on the 40 MHz annular array to allow a more objective comparison to be
made with previously developed arrays.
The pulse-echo responses of the array elements were measured first. A Panametrics
5900PR 200-MHz pulser/receiver (Panametrics Inc., Waltham, MA) was used to excite the
transducer and receive the reflection from a quartz plate placed in a deionized water bath. The
transmit energy and receiver gain for the Panametrics 5900PR were set at 1 μ J and 10 dB,
respectively. RF waveforms were recorded on a digital oscilloscope (LC534 LeCroy Corp.,
Chestnut Ridge, NY) set at 50 Ω coupling. The measured pulse-echo characteristics for all
elements are shown in Table 5.2, and Fig. 5.5 displays the waveform from the element #3. The
average center frequency estimated from the measured pulse-echo responses of array elements
was 38.7 MHz and the −6 dB bandwidth was 51%.
The two-way insertion loss for each element was recorded at 40 MHz. The amplitude of
the sinusoidal signal from an arbitrary function generator (AFG 3251, TektronixInc., Richardson,
TX), set in the burst mode, was measured using 50 Ω coupling on the oscilloscope for various
frequencies over the array’s pass-band. Each array element was then connected to the function
generator with the oscilloscope set at 1 MΩ coupling. The echo signal peak amplitudes for all
eight elements were measured at a distance of 5.3 mm (natural focus for the center element).
Measured data was compensated for loss caused by attenuation in the water bath (2.2 × 10-4
dB/mm.MHz
2
), and transmission coefficient from the quartz target (1.8 dB). The compensated

74

insertion loss values for each element are shown in Table 5.2. The average compensated
insertion loss for the composite array (23.1 dB) was significantly lower than what was reported
for the P(VDF-TrFE)-based annular arrays (31 MHz: 38.4 dB, 55 MHz: 33.5 dB) and 1–3
composite-based array (28 MHz: 32.5 dB) fabricated previously by our group using the same
flexible circuit design (E. J. Gottlieb,2005 and 2006).
The combined electrical and acoustical crosstalk was measured between adjacent
elements. Identical coaxial cables connected array elements to the electronics to provide equal
loading conditions. The annular array was placed in a deionized water bath with no reflector. A
function generator (AFG 3251, Tektronix) set in sinusoid burst mode was used with an
amplitude of 5 Vpp. An element was excited at discrete frequencies through the passband (20 to
60 MHz), and the peak applied voltage was recorded as a reference using an oscilloscope set at 1
MΩ coupling. The voltages on nearest-neighbor elements were also measured with the
oscilloscope set at 1 MΩ coupling and compared with the reference voltage to determine the
level of crosstalk. The results of this test are shown in Fig. 5.6. The maximum combined
crosstalk between the adjacent elements was less than −31 dB over the passband. We attribute
the low crosstalk observed mainly to a non-periodic pillar structure, such as pseudo-random
pillar geometry. These results demonstrate that the pseudo-random pillar composites had
considerably low inter-element cross-talk. Therefore, it is a promising choice for fabricating
imaging annular array ultrasound transducers with low cross-talk.
To show the capabilities of this array for high-frequency ultrasound imaging, images of a
wire phantom will be presented next.

75

Table 5.2 The measured pulse echo characteristics for all annular array elements.

Figure 5.5 Measured time domain pulse echo response (solid line) and normalized frequency
spectrum (dashed line) for element #3.

8.4 8.5 8.6 8.7 8.8 8.9
-300
-200
-100
0
100
200
300
Amplitude ( mV)
Time(us)
10 20 30 40 50 60 70
-36
-30
-24
-18
-12
-6
0
Magnitude (dB)
Frequency(MHz)

Number of Element
1 2 3 4 5 6
Center Frequency
(MHz)
38.7 37.9 38.5 38.1 37.0 38.9
-6 dB Bandwidth
(% C. F.)
50.3 49.2 52.1 51.5 49.6 52.3
-20 dB pulse length
(ns)
92 101 88 93 95 90
Insertion loss (dB)
at 40 MHz
23.3 21.5 24.8 23.2 21.7 23.6

76

Figure 5.6 Cross-talk measured between adjacent elements of the kerfless annular array.

15 20 25 30 35 40 45 50 55 60 65
-60
-55
-50
-45
-40
-35
-30
Amplitude ( dB)
Adjacent elements
1&2
2&3
3&4
4&5
5&6

Frequncy(MHz)

77

5.4 Array Transducer Imaging
Since the lateral interference and spurious modes lead to image artifacts, such as “ghost”
images; to perform a real imaging experiment will be helpful for supplementary determining the
degree of suppression of secondary pulse. This experiment can be conducted by using an
Ultrasound BioMicroscope (UBM) system with a wire phantom, which is common for image
resolution measurement of an ultrasound transducer. A phantom consisting of five 20 µm
diameter tungsten wires (California Fine Wire Co., Grover Beach, CA) was then imaged in order
to assess the axial and lateral resolutions of array transducer. To obtain preliminary images of the
wire target phantom with this array, a synthetic aperture reconstruction algorithm was
implemented. Using this algorithm, responses from each array element can be recorded using
off-the-shelf data acquisition equipment. An excellent review of this imaging technique can be
found in Ritter’s thesis (Ritter, 2001). The interval between the scan-lines was reduced by
increasing PRF to limit the possible scan-line-location error caused by the six independent
measurements. Wire phantom imaging used 2 KHz PRF in 3 µm step sizes between RF lines,
respectively.
The six image frames were summed to form a single frame applying dynamic receive beam
forming. The generated image was displayed using a logarithmic representation of pixel intensity
with a 60 dB dynamic range (see Figure 5.7). The −6 dB lateral resolutions measured at depths
of 3.5, 4, 4.5, 5 and 5.5 mm were 155, 120, 101, 133 and 165 µm, respectively.
An alternative convenient way to estimate resolution, in which all three dimensions are
assessed collectively, is to determine detectability of low-to-high contrast spheres in a speckle
background. Depth ranges of detection are determined; these ranges depend on the sphere

78

diameter and intrinsic contrast. Therefore, to assess performance of high-frequency ultrasound
imagers, a cyst phantom (University of Wisconsin-Madison, Medical Physical Department,
Madison, WI) has been used. It also allow for periodic quality assurance tests and training
technicians in the use of higher-frequency scanners. This phantom contains eight blocks of
tissue-mimicking material; each block contains a spatially random distribution of suitably small
anechoic spheres having a small distribution of diameters. The eight mean sphere diameters are
distributed from 0.10 to 1.09 mm. In this study, only the 1.09 mm mean diameters spheres were
used to determine the performance of the annular array transducer. More details about the design
and fabrication of this cyst phantom were reported elsewhere (Madsen et al. 2010). Figure 5.8
shows measured cyst phantom image with a 50 dB dynamic range.
To quantify improvement, the CNR for the cyst image was calculated. CNR is defined as
the difference between the mean of the background and the cyst in dB divided by the standard
deviation of the background in dB (O’Donnell M. 1988),
tb
b
SS
CNR



(5.1)

where S
t
is the mean of the target, S
b
is the mean of the background, and σ
b
is the standard
deviation of the background. Signals coming from the speckle region are dominated by the main
lobe, thus giving a cross-correlation coefficient near 1. In the case of an anechoic cyst where the
signal contribution from the mainlobe will be small, the sidelobes and grating lobes will be
dominant giving a very low or negative cross-correlation value. The CNR for the cyst phantom
images is 3.3.

79

Overall the fabricated 40 MHz array imaged well considering the limitations of the
monostatic synthetic aperture imaging technique, as well as, physical array imperfections.
Further improvements in image signal to noise ratio, contrast, and reduction of side-lobe artifacts
may be made by using a high frequency multi-channel beamformer for imaging.

80

Figure 5.7 Measured wire phantom image with a 60 dB dynamic range.

Lateral distance [mm]
Axial distance [mm]

-3 -2 -1 0 1 2 3
2
2.5
3
3.5
4
4.5
5
5.5
6
10
20
30
40
50
60

81

Figure 5.8 Measured cyst phantom image with a 50 dB dynamic range.

82

Chapter 6: Summary and Future Work
This study is the first to report the fabrication and use of 1-3 composites with the pseudo-
random pillars to fabricate ultrasound transducers for the purpose of suppressing the lateral
resonances and reducing the crosstalk between array elements. We compared different pillar
geometries and found that the PR composites outperformed the others in specific categories. For
example, the measured electrical impedance of these composites showed that the PR composites
were less affected by the lateral resonance than the SQ and TR composites. In addition,
compared with the SQ and TR composites, the PR composites had the highest relative clamped
permittivity (520 @ 15 MHz) and k
t
values (~ 0.62).
For the PR composites, the level of geometric complexity is determined by the applied
dicing angle sets (25° and 145° were adopted in this paper). There may be a correlation between
different dicing angle sets and the level of crosstalk between array elements; such as 45° and 135°
or 35° and 145°. This could be evaluated in the future in order to determine ways to optimize
the reduction of crosstalk between array elements.
Using these composites, several ultrasound transducers were fabricated to validate their
acoustic performance in water. The measured pulse-echo waveforms and their envelopes showed
that the TR and PR composite transducers both displayed shorter -20 dB pulse lengths when
compared the SQ composite transducer, and the PR composite transducers were the most
sensitive.
Additionally, 15 MHz array sub-apertures with a 0.95 λ pitch were then developed using
these composites to measure their acoustic cross-talk between array elements. Between the
nearest elements, at 15 MHz the PR array sub-aperture had the lowest measured crosstalk, -31.8

83

dB, which was 6.5 dB and 2.2 dB lower than those of the SQ and the TR array sub-apertures,
respectively. Overall, the PR composites outperformed the SQ and TR composites in reducing
the crosstalk between the array elements. A 40 MHz 6-element kerfless annular array transducer
was then designed and fabricated utilizing the PR piezocomposites to reduce inter-element cross-
talk. The average center frequency estimated from the measured pulse-echo responses of array
elements was 38.7 MHz and the −6 dB bandwidth was 51%. The average insertion loss recorded
was 23.1 dB, and the maximum combined crosstalk between the adjacent elements was less than
−31 dB. Therefore, this composite annular array transducer not only provides lower insertion
loss (23.1 dB) than the P(VDF-TrFE)-based annular arrays (E. J. Gottlieb, 2006; 31 MHz: 38.4
dB, 55 MHz: 33.5 dB) and 1–3 composite-based array (28 MHz: 32.5 dB) fabricated previously
by our group using the same flexible circuit design (E. J. Gottlieb, 2005) but also has the lowest
cross-talk. A wire phantom was presented to show the capabilities of this array for high-
frequency ultrasound imaging. ). The −6 dB lateral resolutions measured at depths of 3.5, 4, 4.5,
5 and 5.5 mm were 155, 120, 101, 133 and 165 µm, respectively.
We attribute the lowest crosstalk observed on the PR array sub-aperture mainly to its
randomized composite pillars, which have the highest level of geometric complexity of the three
composites. We hypothesize that the highest level of geometric complexity spreads the acoustic
energy in all lateral directions within the composite; hence, as expected, the PR composite pillars
minimizes the crosstalk between the array elements. Additionally, the PR composites are
advantageous for fabricating high-frequency linear array transducers since they had a higher the
maximum aspect ratio but still performed well. In the future, a study investigating the
relationship between varying the dicing angle set and the crosstalk-reduction may be desirable to
optimize the reduction of the crosstalk between the array elements.

84

Finally, due to their larger pitch and semi-randomized geometry, the PR composites
decreased the fabrication time but still performed well; therefore, when fabricating the high
frequency 1-3 composites, the PR pillar geometry has advantages over the SQ and TR in yield,
convenience, and performance. Considering the observed piezoelectric properties and acoustic
performance, the results presented in this study suggest that 1-3 composites with pseudo-random
pillars may be a better choice for the fabrication of high-frequency piezocomposite single
element and array transducers. However, this work is an initial evaluation of this new composite
geometry and more work, such as modeling with PZFlex is needed for further analysis of its
usefulness.

85

Bibliography
Baer R. L., and Kino G. S., “Theory for cross coupling in ultrasonic transducer arrays,” Appl.
Phys. Lett., vol. 44, pp. 954-956,1984.

Bhatti A. R., Mott M., Evans J. R. G., and Edirisinghe M. J., “PZT pillars for 1–3 composites
prepared by ink-jet printing.” J. Mater. Sci. Lett., vol. 20, pp. 1245–1248, 2001.

Brown J. A., Cherin E., Yin J., Foster F. S., “Fabrication and performance of high-frequency
composite transducers with triangular-pillar geometry,” IEEE Trans. Ultrason., Ferroelect., Freq.
Contr., vol. 56, pp. 827–836, 2009.

Brown J. A., Foster F. S., Needles A., Cherin E., and Lockwood G. R., “Fabrication and
performance of a linear array based on a 1-3 composite with geometric elevation focusing,”
IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 54, no. 9, pp. 1888-1894, 2007.

Cannata J. M. and Shung K. K., “A comparison of model and experiment for a high
frequency(35MHz) linear ultransaonic array,” in Proc. 2003 IEEE Int. Ultrasonics Symp., pp.
1658-1662, 2003.

Cannata J. M., Ritter T. A., Chen W. H., Silverman R. H., and Shung K. K., “Design of efficient,
broadband single-element (20-80 MHz) ultrasonic transducers for medical imaging applications,”
IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 50, no.11, p.p. 1548-1557, 2003.

Cannata J. M., Williams J. A., Zhou Q., Ritter T. A., and Shung K. K., “Development of a 35-
MHz piezo-composite ultrasound array for medical imaging,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr., vol. 53, no. 1, pp. 224-236, 2006.

Cannata J. M., Williams J. A., Zhang L., Hu C. H., and Shung K. K., “A High Frequency Linear
Ultrasonic Array Utilizing an Interdigitally Bonded 2-2 Piezo-Composite,” IEEE Trans.
Ultrason., Ferroelect., Freq. Contr., vol. 58, no. 10, 2011.

Ccrton D., Felix N., Lacaze E., Teston F., and Patat F., “Investigation of Cross-Coupling in 1-3
Piezocomposite Arrays,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 48, no. 1, pp. 85-
92, 2001.

86

Cochran S., Abrar A., Kirk K. J., Zhang D., Button T. W., Su B., Meggs C., and Porch N., “Net-
shape ceramic processing as a route to ultrafine scale 1-3 connectivity piezoelectric ceramic-
polymer composite transducers,” in Proc. 2004 IEEE Int. Ultrasonics Symp., pp. 1682–1685,
2004.

Démoré C. E., Brown J. A., and Lockwood G. R., “Investigation of Cross Talk in Kerfless
Annular Arrays for High-Frequency Imaging,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr.,
vol. 53, no.5, p.p. 1046-1056, 2006.

DeSilets C. S., Transducer Array Suitable for Acoustic Imaging, Ph. D. thesis in Electrical
Engineering, Stanford University, May 1978.

DeSilets C. S., Fraser J. D., and Kino G. S., “Design of efficient broadband piezoelectric
transducers,” IEEE Trans. On Sonics and Ultrasonics, vol. 25, no. 3, pp. 115-125,1978.

Farlow R., Galbraith W., Knowles M., Hayward G., “Micromachining of a Piezocomposite
Transducer using a Copper Vapor Laser,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol.
48, pp.639-640, 2001.

Foster F. S., Ryan L. K., and Turnbull D. H., “Characterization of lead zirconate titanate
ceramics for use in miniature high-frequency (20-80 MHz) transducers, IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr., vol. 38, no. 5, pp. 446-453, 1991.

Foster F. S., Zhang M. Y., Zhou Y. Q., Liu G., Mehi J., Cherin E., Harasiewicz K. A., Starkoski
B. G., Zan L., Knapik D. A., and Adamson S. L., “A new ultrasound instrument for in vivo
microimaging of mice,” Ultras. Med. Biol., vol. 28, no. 9, pp. 1165-1172, 2002.

Foster F. S., Pavlin C. J., Harasiewicz K. A., Christopher D. A., and Turnbull D. H., “Advances
in Ultrasound Microscopy,” Ultras. Med. Biol., vol. 26, no. 1, pp. 1-27, 2000.

Gottlieb E. J., “Development of high frequency annular array ultrasound transducers,” Ph.D.
thesis, Biomedical Engineering, The University of Southern California, Los Angeles, CA, 2005.

Gottlieb E. J., Cannata J. M., Hu C. H., and Shung K. K., “Development of a high-frequency (>
50 MHz) copolymer annular-array, ultrasound transducer,” IEEE Trans. Ultrason., Ferroelect.,
Freq. Contr., vol. 53, no. 5, pp. 1037–1045, 2006.

87

Hackenberger W.S., Kim N., Randall C.A., Cao W., and Shrout T.R., “Processing and
structureproperty relationships for fine grained PZT ceramics,” in Proc. 1996 IEEE Ultrasonics
Symposium, pp. 903–906, 1996.

Hackenberger W., Pan M.J., Kuban D., Ritter T. and Shrout T., “Novel Method for Producing
High Frequency 2-2 Composite from PZT Ceramic”, in Proc. 2000 IEEE Int. Ultrasonics Symp.,
pp. 969-972, 2000.

Hirata Y., Nakaishi H., Numazawa T., and Takada H., “Piezocomposite of fine PZT rods
realized with synchrotron radiation lithography,” in Proc.1997 IEEE Ultrasonics Symp., , pp.
887–881, 1997.

Hamilton E. and Hayward G., “The design of low volume fraction composite transducers using
finite element modeling techniques,” in Proc.1992 IEEE Ultrasonics Symp., pp. 531-534, 1992.

Hayward G.and Bennet J., “Assessing the influence of pillar aspect ratio on the behaviour of 1-3
connectivity composite transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 43,
pp. 98-108, 1996.

Harvey G., Gachagan A., Mackersie J., and Banks R., “Exploring the advantages of a random
1-3 connectivity piezocomposite structure incorporating piezoelectric fibres as the active
element,” in Proc. 2006 IEEE Ultrasonics Symp., pp. 1903-1906, 2006.

Hossack J. A. and Hayward G., “Finite-element analysis of 1-3 composite transducers,” IEEE
Trans. Ultrason., Ferroelect., Freq. Contr., vol. 38, pp. 618-629, 1991.

Hossack J. A., Auld B. A., and Batha H. D., “Techniques for suppressing spurious resonant
modes in 1:3 composite transducers,” in Proc. 1999 IEEE Ultrasonics Symp., pp. 651-655, 1999.

Janas V. F. and Safari A., “Overview of fine-scale piezoelectric ceramic/polymer composite
processing,” J. Amer. Ceram. Soc., vol. 78, pp. 2945–2955, 1995.

Jensen J., “Field: A program for simulating ultrasound systems,” Med. Biol. Eng. Comput., vol.
34, no. 1, pp. 351–353, 1996.

Jensen J., “A new approach to calculating spatial impulse responses,” in Proc. 1997 IEEE
Ultrasonics Symposium, pp. 1755–1760, 1997.

88

Jensen J., and Munk P. “Computer phantoms for simulating ultrasound B-mode and CFM
images,” presented at the 23rd Acoustical imaging symposium, Boston, Massachusetts, 1997.

Krimholtz R, Leedom D. A., and Matthaei G. L., “New equivalent circuit for elementary
piezoelectric transducers,” Electron. Letters, vol. 6, No. 13, pp. 398-399, 1970.

Kwon S., Hackenberger W. et al. “Ceramic/Polymer 2-2 Composites For High Frequency
Tranducers By Tape Casting”, in Proc. 2003 IEEE Int. Ultrasonics Symp., pp. 366-369, 2003.

Lockwood G.R., Ryan L.K., et al., In vitro high resolution imaging in muscular and elastic
arteries,” J. Amer. College Cardiology, vol. 20, pp. 153-160, 1992.

Lockwood G.R., Turnbull D.H., and Foster F.S., “Fabrication of high frequency spherically
shaped ceramic transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 41, no. 2, pp.
231–235, 1994.

Lockwood G.R., Turnbull D.H., Christopher D.A., and Foster F.S., “Beyond 30 MHz,” IEEE
Eng. Med. and Biol., vol. 15, no. 6, pp. 60–71, 1996a.

Lous G. M., Cornejo I. A., McNulty T. F., Safari A., and Danforth S. C., “ Fabrication of curved
ceramic/polymer composite transducer for ultrasonic imaging applications by fused deposition of
ceramics,” in Proc. 1998 IEEE Ultrasonics Symp., pp. 239–242, 1998.

Lukacs M., Yin J., Pang G., Garcia R. C., Cherin E., Williams R., Mehi J., and Foster F. S.,
“Performance and characterization of new micromachined high-frequency linear arrays,” IEEE
Trans. Ultrason., Ferroelect., Freq. Contr., vol. 53, no. 10, pp. 1719–1729, 2006.

Liu R., Harasiewicz K. A., Knapik D., Freeman N. A., and Foster F. S., “2-2 piezoelectric
composites with high density and fine scale fabricated by interdigital pair bonding,” Appl. Phys.
Lett., vol. 75, pp. 3390-3392, 1999.

Liu R., Harasiewicz K. A., and Foster F. S., “Interdigital pair bonding for high frequency (20–50
MHz) ultrasonic composite transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol.
48, no. 1, pp. 299–306, 2001.

Madsen E. L., Frank G. R., Mccormick M. M., Deaner M. E., and Stiles T. A., “Anechoic sphere
phantoms for estimating 3-D resolution of very-high-frequency ultrasound scanners.” IEEE
Trans. Ultrason., Ferroelect., Freq. Contr., vol. 57, no. 10, pp. 2284–2292, 2010.

89

Michau S., Mauchamp P., and Dufait R., “ Piezocomposite 30 MHz linear array for medical
imaging: Design challenges and performances evaluation of a 128 elements array,” in Proc.2004
IEEE Ultrasonics Symp., pp. 898–901,2004.

Newnham R. E., Skinner D. P., and Cross L. E., “Connectivity and piezoelectric-pyroelectric
composites,” Mat. Res. Bull., vol. 13, pp. 525–536, 1978.

Noguera R., Lejeune M., Chartier T., “3D fine scale ceramic components formed by ink-jet
prototyping process,” Journal of the European Ceramic Society, vol. 25, pp. 2055–2059, 2005.

Oakley C.G., “Single crystal piezoelectric: a revolutionary development for transducers,” in Proc.
2000 IEEE Ultrasonics Symposium, vol. 2, pp. 1157–1167, 2000.
O’Donnell M., and Flax S. W., “Phase-aberration correction using signals from point reflectors
and diffuse scatterers: Measurements,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 35,
no. 6, pp. 768–774, 1988.

Pang Guofeng, Sayer Micheal and Lockwood G. R., “Fabrication of PZT Sol Gel Composite
Ultrasonic Transducers Using Batch Fabrication Micromolding,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr., vol. 53, pp. 1679-1684, 2006.

Park S.E., and Shrout T.R., “Characteristics of relaxor-based piezoelectric single crystals for
ultrasonic transducers,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 44, no. 5, pp.
1140–1147, 1997.

Passman C., Ermert H., Auer T., Kasper K., Gammal E., and Altmeyer P., “ In vivo ultrasound
biomicroscopy,” in Proc. 1993 IEEE Ultrasonics Symp., pp. 1015-1018, 1993.

Ritter T.A., Geng X., Shung K.K., Lopath P.D., Park S-E., and Shrout T.R., “Single crystal
PZN/PT-polymer composites for ultrasound transducer applications,” IEEE Trans. Ultrason.,
Ferroelect., Freq. Contr.,, vol. 47, no. 4, pp. 792–800, 2000.

Ritter T.A., Design, Fabrication, and Testing of High Frequency (>20 MHz) Composite
Ultrasound Imaging Arrays, Ph.D. thesis in Bioengineering, The Pennsylvania State University,
Aug. 2000.

Ritter T.A., Shrout T.R., and Shung K.K., “A high frequency ultrasound array incorporating an
actuator,” Proc. Med. Imaging 2001: Ultrason. Transducer Eng, vol. 4325, pp. 36-46, 2001.

90

Ritter T. A., Shrout T. R., Tutwiler R., and Shung K. K., “A 30-MHz piezo-composite ultrasound
array for medical imaging applications,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol.
49, pp.217-230, 2002.

Reynolds P., Hyslop J., and Hayward G., “Analysis of spurious resonances in single and multi-
element piezocomposite ultrasonic transducers,” in Proc. 2003 IEEE Ultrasonics Symp., , pp.
1650–1653, 2003.

Savakus H. P., Klicker K. A., and Newnham R. E., “PZTEpoxy piezoelectric transducer: A
simplified fabrication procedure,” Mat. Res. Bull., vol. 16, pp. 677–680, 1981.

Shung K. K., Smith M. B., and Tsui B. M. W., Principles of Medical Imaging, San Diego, CA:
Academic Press, 1992.

Shung K. K., Diagnostic Ultrasound: Imaging and Blood Flow Measurements, Boca Raton, FL:
Taylor & Francis Group, LLC, 2006.

Shrout T., “ Innovations in piezoelectric materials for ultrasound transducers,” in Proc. 2008
IEEE Int. ultrasonics Symp., vol. 2, 2008.

Silverman R.H., Lizzi F. L., Ursea B. G., Cozzarelli L., Ketterling J. A., Deng C. X., Folberg R.,
and Colemen D. J., “Safety levels for exposure of cornea and lens to very high-frequency
ultrasound,” J. Ultras. Med., vol. 20, pp. 979-986, 2001.

Smith W. A., “The role of piezocomposites in ultrasonic transducers,” in Proc. 1989 IEEE Int.
Ultrasonics Symp., pp. 755–766, 1989.

Smith W. A. and B. A. Auld, “Modeling 1-3 composite piezoelectrics: Thickness-mode
oscillations,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 38, pp. 40-47, 1991.

Smith W. A., “New opportunities in ultrasonic transducers emerging from innovations in
piezoelectric materails,” in Proc. SPIE Int. Symp., 1992, vol.1733, pp.3-26.

Turnbull D. H., Starkoski B. G., Harasiewicz K. A., Semple J. L., From L., Gupta A. K., Sauder
D. N., and Foster F. S., “A 40-100 MHz B-scan ultrasound backscatter microscope for skin
imaging,” Ultrason. Med. Biol., vol. 21, no. 1, pp. 79-88, 1995.

91

Turnbull D. H., “In utero ultrasound backscatter microscopy of early stage mouse embryos,”
Comput. Med. Imaging Graphics, vol. 23, no. 1, pp. 25-31, 1999.

Wang S., Li X., Wakabayashi K., Esashi M., “Deep reactive ion etching of lead zirconate titanate
using sulfur hexafluoride gas,” J. Am. Ceram. Soc., vol. 82, pp. 1339-1341, 1999.

Wang S., Li J.F., Wakabayashi K., Esashi M., Watanabe R., “Lost Silicon Mold Process for PZT
Microstructures”, Adv. Mater. vol. 11, pp. 873-876, 1999.

Wang H., Ritter T. A., Cao W., and Shung K. K., "High frequency properties of passive
materials for ultrasonic transducers", IEEE Trans. Ultrason., Ferroelect., Freq. Contr., vol. 48,
no. 1, pp. 78-84, 2001.

Yin J., Lee M., Brown J. A., Cherin E., and Foster F. S., “Effect of Triangular Pillar Geometry
on High-Frequency Piezocomposite Transducers,” IEEE Trans. Ultrason., Ferroelect., Freq.
Contr., vol. 57, No. 4, pp. 957–968, 2010.

Yin J., Lukacs M., Harasiewicz K. A., and Foster F. S., “Design and fabrication of ultrafine
piezoelectric composites,” Ultrason. Imaging, vol. 27, no. 1, pp. 54-64, 2005.

Yu D.S., Han J.C., Ba L., “PbTiO3 nanosized ceramics,” American Ceramic Society Bulletin, vol.
81, No. 9, pp 38-39, 2002.

Yuan J., Marsh P., Liang K., and Kunkel H., “The dynamic characteristics of composites with
periodic and random structure,” in Proc. 1996 IEEE Ultrasonics Symp., pp. 949-954, 1996.

Yuan J. R., Jiang X., Cao P. J., Sadaka A., Bautista R., Snook K., and Rehrig P. W., “High
frequency piezo composites microfabricated ultrasound transducers for intravascular imaging,”
in Proc. 2006 IEEE Ultrasonics Symp., pp. 264–268, 2006.

92

Yuan J. R., et al., “Microfabrication of piezoelectric composite ultrasound transducers (PC-
MUT).” in Proc. 2006 IEEE Int. Ultrasonics Symp., pp. 918–921, 2006.

Zhao J. Z., Alves C. H. F., Snook K. A., Cannata J. M., Chen W. H., Meyer R. J., Ayyappan Jr.,
S., Ritter T. A., and Shung K. K., “ Performance of 50 MHz transducers incorporating fiber
composite PVDF, PbTiO3 and LiNbO3,” in Proc. 1999 IEEE Ultrasonics Symp., pp. 1185–1189,
1999.

Abstract (if available)

Abstract The goal of this research was to develop a novel diced 1-3 piezocomposite geometry to reduce pulse-echo ring down and acoustic cross-talk between high-frequency ultrasonic array elements. PZT-5H based 1-3 composites (10 and 15 MHz) of different pillar geometries (square (SQ), 45° triangle (TR), and pseudo-random (PR)) were fabricated and then made into single element ultrasound transducers. The measured pulse-echo waveforms and their envelopes indicate that the PR composites had the shortest -20 dB pulse length and highest sensitivity among the composites evaluated. Using these composites, 15 MHz array sub-apertures (0.95 λ pitch) were fabricated to assess the acoustic cross-talk between array elements. The combined electrical and acoustical cross-talk between the nearest array elements of the PR array sub-apertures (-31.8 dB @ 15 MHz) was 6.5 dB and 2.2 dB lower than that of the SQ and the TR array sub-apertures, respectively. To further reduce inter-element cross-talk of kerfless annular array transducers, utilizing the piezocomposites with the PR pillar geometry, we fabricated high-frequency annular arrays. Each annular array was designed to have six equal-area elements and a center frequency of 40 MHz. The average center frequency estimated from the measured pulse-echo responses of array elements was 38.7 MHz and the −6 dB bandwidth was 51%. The average insertion loss recorded was 23.1 dB, and the maximum combined crosstalk between the adjacent elements was less than −31 dB. These results demonstrate that the 1-3 piezocomposite with the pseudo-random pillars may be a better choice for fabricating enhanced high frequency linear array ultrasound transducers

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Array transducers for high frequency ultrasound imaging

PDF

Development of high frequency focused transducers for single beam acoustic tweezers

PDF

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

PDF

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

PDF

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

PDF

High frequency ultrasonic phased array system and its applications

PDF

Development of front-end circuits for high frequency ultrasound system

PDF

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

PDF

Intravascular imaging on high-frequency ultrasound combined with optical modalities

PDF

Highly integrated 2D ultrasonic arrays and electronics for modular large apertures

PDF

Configurable imaging platform for super-harmonic contrast-enhanced ultrasound imaging

PDF

A high frequency array- based photoacoustic microscopy imaging system

PDF

High frequency ultrasound array for ultrasound-guided breast biopsy

PDF

High-frequency ultrasonic transducers for photoacoustic applications

PDF

Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications

PDF

Design and development of ultrasonic array transducers for specialized applications

PDF

Fabrication of ultrasound transducer and 3D-prinitng ultrasonic device

PDF

Single-cell analysis with high frequency ultrasound

PDF

Microfluidic cell sorting with a high frequency ultrasound beam

PDF

Development of back-end processing system for high frequency ultrasound b-mode imaging

Asset Metadata

Creator Yang, Hao-Chung (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Publication Date 07/30/2012

Defense Date 05/24/2012

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

Tag crosstalk,OAI-PMH Harvest,piezocomposites,pseudo-random,transducers,ultrasound

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Goo, Edward K. (committee member), Nutt, Steven R. (committee member), Shung, Kirk Koping (committee member), Yen, Jesse T. (committee member)

Creator Email yangh@usc.edu,Yanghaochung@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-77190

Unique identifier UC11289972

Identifier usctheses-c3-77190 (legacy record id)

Legacy Identifier etd-YangHaoChu-1066.pdf

Dmrecord 77190

Document Type Dissertation

Rights Yang, Hao-Chung

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA