University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

2D Ultrasonic Transducer Array’s Design and Fabrication with 3D Printed Interposer and
Applications
by

Haochen Kang

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
(BIOMEDICAL ENGINEERING)

August 2022
Copyright © 2022 Haochen Kang
ii
Dedication

I dedicate this thesis to my family for their selfless dedication and love.
Without your support, this journey would be impossible for me.

iii

Acknowledgments

It always takes a great team to face all the challenges in projects. I wish to
acknowledge the tremendous help and support from the people who made this
possible.
First, I would like to thank my advisor Prof. Qifa Zhou for his years of guidance and
instruction and for all the resources and assistance he has provided, which have been
the cornerstone of all my work. I would also like to acknowledge the help from Prof.
Zhongping Chen from the Department of Biomedical Engineering and the Beckman
Laser Institute, University of California, Irvine.
Over the years of working in the lab, I have built up a deep friendship with the lab
members. Good teamwork has become an essential help for every project and the
friendship and the good atmosphere in our have has been the key to maintaining my
optimism in the face of challenges.
In particular, I wish to give thanks to Dr. Robert Wodnicki who has been both a great
mentor and great friend. His passion and experience in research have inspired and
helped me in my work. I would also like to acknowledge Nestor Cabrera-Munoz, Ruimin
Chen, Yizhe Sun, and Yushun Zeng for their tremendous help in transducer fabrication
and testing. It takes years of work and practice to discover an effective production
process. Xuejun Qian, Runze Li, and Fengyi Zhang made great contributions to

iv

elastography relates works and animal studying. From Payam Eliahoo, Gengxi Lu,
Laiming Jiang, and Zeyu Chen, I learned valuable knowledge about circuit design,
piezoelectric materials, and acoustic physics during discussions with them. And my
special thanks to Yansong Zhu, a close friend for many years, who provided me with
precious knowledge about image processing, especially the super-resolution imaging
technology.
Last but not least, I wish to express my gratitude to my family. Especially in such
special times, their support and encouragement are the most important force for me to
keep going.

v

Table of Contents
Dedication ............................................................................................................................ii
Acknowledgments............................................................................................................... iii
List of Tables ...................................................................................................................... vii
List of Figures .................................................................................................................... viii
Abstract ............................................................................................................................... xi
Chapter 1 Instruction .......................................................................................................... 1
1.1 Background:............................................................................................................... 1
1.1.1 2D Ultrasonic Transducer Array ......................................................................... 1
1.1.2 Acoustic Radiation Force Optical Coherent Elastography ................................ 11
1.1.3 Stereolithography 3D printing technology ....................................................... 15
1.2 Motivations and Objectives .................................................................................... 18
1.3 Outline ..................................................................................................................... 20
Chapter 2 2D Ultrasonic Ring Array Fabrication with 3D printed interposer ................... 23
2.1 Array Design ............................................................................................................ 23
2.2 Field II Simulation .................................................................................................... 26
2.3 KLM model-based simulation .................................................................................. 31
2.4 3D printed interposer design and filling ................................................................. 34
2.5 Array assembly and fabrication............................................................................... 38
2.6 Array test ................................................................................................................. 45
2.7 Discussion ................................................................................................................ 54
Chapter 3 2D array based OCE system Experiments results ............................................ 58
3.1 Systems setup and imaging process........................................................................ 58
3.2 Phantom Test and Ex-vivo Tissue Results ............................................................... 60
3.2.1 G1 ring array’s Phantom Results ...................................................................... 60
3.2.2 G2 ring array’s Phantom Results ...................................................................... 62
3.2.2 G1 ring array’s Ex-vivo tissue results ................................................................ 65
3.2.2 G2 ring array’s Ex-vivo tissue results ................................................................ 67
3.3 G2 ring array’s In-vivo tissue results ....................................................................... 70
3.4 Discussion ................................................................................................................ 72
Chapter 4 10MHz 2D array based on the pitch-shifting interposer ................................. 77
4.1 2D array and pitch-shifting interposer design ........................................................ 77
4.2 Circuit design ........................................................................................................... 80
4.3 Interposer fabrication ............................................................................................. 81

vi

4.4 Array fabrication...................................................................................................... 84
4.5 Performance testing and imaging results ............................................................... 86
4.6 Discussion ................................................................................................................ 91
Chapter 5 Summary and Future Work .............................................................................. 94
5.1 Summary and discussion ......................................................................................... 94
5.2 Future Work ............................................................................................................ 97
BIBLIOGRAPHY ................................................................................................................ 102

vii

List of Tables
TABLE 2-1 .......................................................................................................................... 24
TABLE 2-2 .......................................................................................................................... 25
TABLE 4-1 .......................................................................................................................... 77

viii

List of Figures
Fig. 1-1. ................................................................................................................................ 1
Fig. 1-2. ................................................................................................................................ 2
Fig. 1-3. ................................................................................................................................ 4
Fig. 1-4. ................................................................................................................................ 6
Fig. 1-5. ................................................................................................................................ 7
Fig. 1-6. ................................................................................................................................ 8
Fig. 1-7. ................................................................................................................................ 9
Fig. 1-8. .............................................................................................................................. 10
Fig. 1-9. .............................................................................................................................. 14
Fig. 1-10. ............................................................................................................................ 16
Fig. 1-11. ............................................................................................................................ 19
Fig. 2-1. .............................................................................................................................. 23
Fig. 2-2. .............................................................................................................................. 27
Fig. 2-3. .............................................................................................................................. 28
Fig. 2-4. .............................................................................................................................. 28
Fig. 2-5. .............................................................................................................................. 29
Fig. 2-6. .............................................................................................................................. 30
Fig. 2-7. .............................................................................................................................. 31
Fig. 2-8. .............................................................................................................................. 32
Fig. 2-9. .............................................................................................................................. 33

ix

Fig. 2-10. ............................................................................................................................ 34
Fig. 2-11. ............................................................................................................................ 36
Fig. 2-12. ............................................................................................................................ 39
Fig. 2-13. ............................................................................................................................ 40
Fig. 2-14. ............................................................................................................................ 42
Fig. 2-15. ............................................................................................................................ 43
Fig. 2-16. ............................................................................................................................ 44
Fig. 2-17. ............................................................................................................................ 45
Fig. 2-18. ............................................................................................................................ 45
Fig. 2-19. ............................................................................................................................ 46
Fig. 2-20. ............................................................................................................................ 47
Fig. 2-21. ............................................................................................................................ 49
Fig. 2-22. ............................................................................................................................ 49
Fig. 2-23. ............................................................................................................................ 51
Fig. 2-24. ............................................................................................................................ 52
Fig. 2-25. ............................................................................................................................ 53
Fig. 2-26. ............................................................................................................................ 53
Fig. 3-1. .............................................................................................................................. 59
Fig. 3-2. .............................................................................................................................. 59
Fig. 3-3. .............................................................................................................................. 61
Fig. 3-4. .............................................................................................................................. 63
Fig. 3-5. .............................................................................................................................. 64

x

Fig. 3-6. .............................................................................................................................. 66
Fig. 3-7. .............................................................................................................................. 68
Fig. 3-8. .............................................................................................................................. 69
Fig. 3-9. .............................................................................................................................. 71
Fig. 4-1. .............................................................................................................................. 78
Fig. 4-2. .............................................................................................................................. 79
Fig. 4-3. .............................................................................................................................. 81
Fig. 4-4. .............................................................................................................................. 82
Fig. 4-5. .............................................................................................................................. 84
Fig. 4-6. .............................................................................................................................. 85
Fig. 4-7. .............................................................................................................................. 87
Fig. 4-8. .............................................................................................................................. 87
Fig. 4-9. .............................................................................................................................. 88
Fig. 4-10. ............................................................................................................................ 89
Fig. 4-11. ............................................................................................................................ 90
Fig. 5-1. .............................................................................................................................. 96
Fig. 5-2. .............................................................................................................................. 96
Fig. 5-3. .............................................................................................................................. 98
Fig. 5-4. .............................................................................................................................. 99

xi

Abstract
2D array fabrication has been an exciting research direction in the past 30 years.
Connecting high density element matrix to the system has been one of the main
challengings that limit the development of a fully sampled 2D array. The rapid
development in 3D printing technology in recent years provides an option to address
this challenge.
In this work, we implemented 3D printed interposers with straight channels and
progressed to interposers with pitch-shifting structures in fully sampled 2D array
fabrication and designed and built different types of 2D arrays targeted at different
applications: 2D array with central opening geometry for acoustic radiation force optical
coherence elastography (ARF-OCE) and high-frequency (>10MHz) imaging array. ARF-
OCE has been successfully implemented to characterize the biomechanical properties of
soft tissues such as the cornea and the retina with high resolution using single-element
ultrasonic transducers for ARF excitation. In this study, we combined the advantages of
3D dynamic electronic steering of the 2D ultrasonic array and high-resolution optical
coherence tomography (OCT) and propose a new method called 2D ultrasonic array-
based optical coherence elastography imaging. And 2D imaging array, which enabled
real-time volumetric imaging, has shown its great potential in transesophageal
echocardiography (TEE) and intracardiac echocardiography (ICE). The performance of
the 2D arrays’ elements was measured in pulse-echo(P/E) test and the Ring array’s 3D
steering capability was first validated using a hydrophone. The combined 2D ultrasonic

xii

array OCE system was calibrated using a homogenous phantom, followed by an
experiment on ex vivo rabbit corneal tissue. And the high-frequency 2D array was
applied in the imaging test and real-time 3D images were acquired. The results
demonstrate that the 3D printed interpose offers another possibility to address the
limitation of high-density interconnection and strict geometry requirements in 2D array
design and fabrication.

1

Chapter 1 Instruction
1.1 Background:
1.1.1 2D Ultrasonic Transducer Array
Ultrasound transducer arrays are playing an important role in modern medical imaging.
Different types of ultrasound arrays were designed based on the application scenarios. Up until
now, one-dimensional arrays are still the mainstream products in the market. The classification
of the types is usually determined by their size and shape, and can be grouped into the
following categories (fig. 1-1): Linear array, phased array, and Curvilinear arrays [1, 2]. Linear
arrays, with one wavelength (at the central frequency of the array) element pitch, usually

Fig. 1-1. Brief illustrations of the different probe types for acquiring B-mode images [1].

2

provide a rectangular field of view. This type of array is commonly used to scan near-surface
anatomies such as breast, superficial musculoskeletal, and kidney [2, 3]. Phased arrays are
designed with wavelength (at the central frequency of the array) element pitch and are
primarily used for cardiac imaging [4]. Benefiting from the fine pitch, the phased array’s beam
can be steered up to ±45° [2]. Compared with the linear array (maximum steering angle is
usually limited to ±15°) [2], the phased array provides a wider field of view with a narrower
scanning window. Curvilinear array, to some degree, is a combination of the previous two. One
wavelength pitch design with a curve shape ensures a large field of view and deep penetration.
It is widely used for abdominal imaging [2]. With 1D arrays, volumetric images can be acquired
with freehand or mechanically scanning (Fig. 1-2) [5-10]. Unfortunately, the frame rate of this

Fig. 1-2. Examples of freehand 3D ultrasound imaging and reconstruction systems [5]: (a) Freehand 3D imaging system with an
optical tracking system [6]. (b) Freehand 3D imaging system based on a 3D mechanical probe ultrasound transducer [7]. (c)
Freehand 3D imaging system using a mechanical external fixture [8].

3

approach was significantly limited [2, 11], which will further influence the image quality and
limits its imaging methods. Thus, just like the transition from the scanning single element to 1D
arrays, to achieve real-time volumetric imaging, a 2D ultrasound array was presented [12].
In recent years, 2D arrays have been applied in clinical usage and their application was
mainly focused on echocardiography and fetal imaging [2, 13]. With a matrix of thousand
elements, these 2D arrays provide real-time 3D imaging and become very helpful tools for
physicians. Besides, many advanced research fields for ultrasound application could benefit
from the development of 2D arrays. For example, 3D super-resolution imaging (SRI) technology.
By localizing the microbubbles in multiple frames, the trace of the bubbles in blood flow will be
yielded, which represents the structure of the microvessels. With this technology, the
microvessels whose scales are below the diffraction limits of the acoustic wave, which provides
the theoretical minimum value of the structure size for ultrasound imaging, can be mapped.
Even with the probes operated at a relatively low-frequency range, the resolution of the images
can reach a few micrometers. Numerous research on SRI has been reported in recent years.
Errico, Claudia, et al [14] presented in vivo SRI results of a rat brain in 2D by implementing
ultrafast ultrasound localization microscopy(ULM), and the in-plane blood flow velocities were
acquired. It is conceivable that the extension of this technology from 2D images to 3D images
could bring great benefits to the field of microvascular imaging. Several researches have been
presented to perform SRI technology with mechanically scanning 1D array [14, 15]. Yet this
setup failed in estimating the out-of-plane location and suffered from a long acquisition time
[16]. In 1D scanning, only the information within the field of view will be recorded. The data of

4

blood flow that is orthogonal to the imaging plane is missing and the bubbles are not able to be
traced. Besides, collecting enough separable sources is necessary for ULM. Mechanically
scanning different planes and combining them to form volumetric imaging requires a
considerably long acquisition time. Considering the lifetime of the microbubbles and the
requirement of the concentration of the bubbles for SRI application, this will lead to further
problems in in-vivo applications. Thus, 2D arrays, which realize real-time 3D imaging, will be

Fig. 1-3. 3D super-resolution images with the sparse 2D array. To acquire the SRI results, microbubble solution was injected into
the tubes and flowed in the opposite direction. Top: 3D super-resolution image of two tubes. The photo at the top right corner
presents the target; Bottom: Velocity maps [17].

5

essential components of future 3D super-resolution studies to overcome the weaknesses and
limitations that exist in the approach with 1D arrays. Recently, promising results have been
reported [16, 17]. With the 2D array, the flow in 3D space can be reconstructed and the velocity
of the flow can be measured (fig. 1-3).
With these benefits, research on 2D array fabrication has been a hugely popular area in the
past 30 years. Tremendous development has been achieved in matrix ultrasound transducer
array fabrication and application in the past few decades. Unfortunately, the difficulties of
fabrication impede the further development of 2D ultrasound array technology. The
conventional fabrication process for 1D arrays is difficult to apply for a 2D array due to its high
element density. Fiering et al [18], presented a customized flexible printed circuit (FPC) with a
complex fabrication process that was implemented to build a 2D array. Up until today, the
capability of most FPC manufacturers is still hard to meet the requirement of the fabrication for
a 2D array with an operation frequency higher than 5MHz. One of the main challenges in 2D
array fabrication is the interconnection of the elements to the driving system. From 1D array to
2D array, the size of the elements and aperture area is reducing, while the total element
number is increasing. The result is the element density, the total elements per unit area is
increasing expensively. In the face of such challenges, multiple solutions have been provided to
relieve this bottleneck in the past few decades. Since FPC’s trace and space could not support
the high demands of 2D array fabrication, researchers naturally turn to the chip manufacturing
industry to respond to the 2D array's demanding circuit design requirements with the help of
semiconductor technology. Thus, application-specific integrated circuits (ASICs) have become

6

one of the most common solutions in facing the challenges in 2D array fabrication. The AISCs
designed with a pre-amplifier, multiplexer, or micro-beamformer for sub-array processing were
designed for fully-sampled matrix arrays with the purpose to reduce the channel number to a
more reasonable value for most ultrasound systems [19-22]. Such ASIC-based matrix arrays
have been commonly fabricated directly on the surface of the IC substrate itself, which benefits
the signal-to-noise ratio in addition to relieving the 2D routing bottleneck. Thus, in such pitch-
matched designs, the ASICs’ pads and related circuitry need to fit into the area of an element.
With the help of semiconductor technology, great research has been done in academia and

Fig. 1-4. ASIC embedded with 3 × 3 sub-array receivers for 2D array fabrication [19]. The 2D array was directly assembled on the
ASIC. a. Overview of the ASIC; b. Floor plan of a 3 × 3 sub-array receiver in the ASIC; c. A 32 by 32 PZT array integrated with the
ASIC; d. ASIC bonded on the PCB for the experiment.

7

industry and presents different ASICs with promising performance. Presented by Chen et al
[19], a 32 × 32 element 2D array with separated transmission and receive channels were
fabricated with the ASIC implemented with a 3 by 3 sub-array receiver design [Fig. 1-4], and a
larger aperture of 1.75D array based on ASICs with MUX was reported by Wodnicki et al [20].
Even though these designs represent significant work for trading off design parameters to fit
with array requirements and improve performance, the limited area still constrains the
performance and becomes one of the main challenges for the ASIC design. The complexity of
these tradeoffs is further compounded by the fact that in ultrasound it is necessary to utilize
both low voltage and high voltage ASIC fabrication processes. These situations and

Fig. 1-5. Illustration of different array layouts. The blue block and red block represent transmit element and receive elements
respectively [24]. Elements in purple means they work in both transmission and receiving. a. fully sampled array; b-f: sparse
array with a different algorithm.

8

requirements make ASIC's design increasingly tricky and challenging, especially when the
operation frequency of the transducer aims at a relatively high region, e.g., 10 MHz.
ASICs are not the only methods to address the challenges in 2D array fabrication. Aiming at
reducing the density of elements, sparse array [17, 23-26] and row-column array [16, 27-29]
were brought forward. A sparse 2D array makes a trade-off between the grating lobes and the
active element numbers [30, 31]. Different from the conventional fully sampled array, the
elements in a sparse array are not necessarily spaced closely to each other in the aperture
matrix. It does not even follow a periodic or regular order of arrangement [Fig. 1-5, 1-6]. Thus,
the element density in a sparse array will be significantly reduced and the aperture size will be
expanded compared to a fully sampled array with the same number of active elements.

Fig. 1-6. The layout of a 2D sparse array is applied for 3D SRI [17]. A typical optimized rectangular 2D sparse array.

9

Randomly placing the elements is a reliable approach to building a sparse array. Multiple
optimization theories aimed at improving image quality and reducing side lobes were proposed
in recent years [26, 32, 33]. Meanwhile, as another promising research direction for sparse
array image quality improvement, non-rectangular 2D sparse arrays, such as spiral array and
concentric circular arrays with irregular geometry, is also being studied extensively [24, 34-36].
The row-column array is another popular design to reduce the complexity of the 2D array
fabrication, it provides deep penetration depth and a high-volume rate. Its ability in 3D imaging
has been presented by several groups [28, 29, 37]. In a row-column array, the arrangement
direction of the electrodes on the top and bottom side of the elements are orthogonal [Fig. 1-
7], which gives the row-column array the ability to scan in two directions and acquire the
information from 3D space. Recently, a forward-looking miniature endoscope designed for 3D

Fig. 1-7. Illustration of row-column array layout [28]. The geometry of the element matrix is similar to a fully sampled array,
except the elements are shorted in different directions by the electrodes on the top and bottom surface, which ends up in
row element array and column element array.

10

imaging was presented by Katherine Latham et al [29]. This device was based on a 30-MHz row-
column array with 64 × 64 orthogonal elements. Utilizing a previously reported novel imaging
technique [38], this array presented a clear 3D image of two crossed, depth offset wires with
good resolution. However, the novel imaging technique requires different pulse polarity on
transmit and receiving which presents further challenges in circuit and system design.
Besides the relatively mature work introduced above. The studies on further reducing the
number of elements in the 2D array while maintaining the image quality present more
possibilities. For example, the orthogonal array and boundary array presented by Yen et al [12,
39].
In previous work, a mature 2D array fabrication process based a on 3D printed interposer
was developed. The interposer implemented in the 2D array served as a backing layer as well as
an interconnection part that solved the issue in the strong ring down when 2D arrays were
directly assembled on ASIC or PCB [Fig. 1-8]. The conventional process, in which the transducer
arrays were directly attached to a flexible printed circuit, implemented in linear array

Fig. 1-8. Interposer based 2D array fabrication. a. 2D array fabricated on interposer; b. cross-section of the composite and the
interposer channels; c. 2D array with PCB after assembly. [20]

11

fabrication will lead to several issues in 2D array fabrication. ASIC or PCB with multiple layers
cannot serve as a good backing block in 2D array fabrication, which will lead to strong
ringdown. In interposers, the channels filled with conductive epoxy supported the backside of
each element. They connected the element to the electrodes on the circuit as well as absorbed
the acoustic wave on the back side. In those works, the interposers were designed with straight
channels [Fig. 1-8. a, b]. The geometry of the channel matrix maintains the same on the top and
bottom sides. Thus, with these interposers, the limitation of the scale still exists in circuit
design. Limited by the available 3D printing technologies during those projects, the potential of
special architectures in interposers was not further explored and discussed. Nevertheless, the
idea of interposers with a special channel structure was formed.
In the following chapters, the details of the term “pitch-shifting interposer” will be
introduced. With the development of 3D printing technologies, this special interposer structure
design became possible, which makes 2D arrays very affordable and available for different
applications, such as 2D array based elastography.
1.1.2 Acoustic Radiation Force Optical Coherent Elastography
Aging, disease, and other factors lead to changes in the biomechanical properties of the eye
[40, 41]. Glaucoma, an optic neuropathy that is one of the main causes of irreversible blindness
worldwide [42], has been shown to have a strong relationship with changes in the
biomechanical properties of corneal tissue [40, 41, 43, 44]. Recent research predicts that there
will be 80 million people suffering from glaucoma and over 3.2 million people blinded due to
this disease burden [45]. Mechanical property changes in the cornea also carry important

12

implications for medical interventions for the eye such as laser in situ keratomileuses (LASIK)
[46]. LASIK procedures may cause a reduction in biomechanical integrity of the cornea and carry
a potential risk for keratoconus, an ecstatic corneal disorder, and iatrogenic keratectasia, a
complication after LASIK that may compromise the integrity of the posterior stromal bed. [47-
49] Optical Coherence Tomography (OCT), a recently developed high-resolution imaging
modality, has been widely adopted in the clinic to assist ophthalmologists in the assessment of
ocular disease. In recent research, Spectral Domain (SD)-OCT has been used for glaucoma
diagnosis by measuring the average retinal nerve fiber layer (RNFL) thickness [50]. However,
the ability of SD-OCT to discriminate between healthy and glaucomatous eyes is limited[50].
Even though OCT imaging provides crucial information for the diagnosis of common eye
conditions, its strength is limited in cases where the disease is difficult to detect from subtle
morphological changes only, especially at the early stages of disease progression.
Another example is keratoconus, for which the corneal biomechanical properties are
important for early diagnosis [51]. Therefore, methods that can provide biomechanical
properties at micrometer resolution hold significant promise for use in the clinical setting for
diagnosis, as well as for the selection of therapies [52].
Different techniques, such as ocular response analyzer (ORA) and noncontact tonometry
(Corvis ST, Oculus, Wetzlar, Germany), have been employed in clinical practice for measuring
the biomechanical properties of the cornea. These techniques analyze the response of the
tissues to an air pulse to measure the tissues’ biomechanical properties [53, 54]. However,
these methods can only be used to measure the average biomechanical properties of a large

13

tissue area [40]. In addition, the air puff technique can generate large-amplitude deformations
in tissue, which leads to low accuracy results [52]. These issues limit the application of these
techniques when mapping and detecting subtle stiffness differences of ocular diseases at an
early stage. Elastography is an imaging method with the capability to yield quantitative probing
of biomechanical properties of soft tissue [55], making it a powerful tool for diagnosing
diseases that lead to tissue elasticity changes such as cancer [56, 57] and fatty liver [58].
However, current clinically available elastography techniques such as ultrasound elastography
[59, 60] and magnetic resonance elastography, do not produce high spatial resolution results
and are therefore inapplicable to ocular tissues which have thicknesses in the range of a few
hundred microns [61-64]. Optical Coherence Elastography (OCE), utilizing high-resolution OCT
to detect the propagation of induced mechanical waves, has been recently developed to meet
the requirements for ophthalmologic applications [Fig. 1-9]. The bandwidth and wavelength of
generated mechanical waves for OCE depend on the temporal and spatial characteristics of the
excitation push [40]. Currently utilized pushing sources for OCE imaging include either air-puff
or pulsed laser excitation [40, 65, 66]. However, the air-puff method has certain limitations
such as low bandwidth and low repetition rate while the pulsed laser approach has safety
limitations. Compared with these pushing methods, acoustic radiation force (ARF) is non-
contact, steerable in 2D with a matrix array, and can control the spatiotemporal characteristics
of the induced mechanical waves, making it a promising pushing source for OCE.

14

In the current acoustic radiation force optical coherence elastography (ARF-OCE) approach,
single-element-based ultrasonic transducers are typically used [67-69]. However, some
disadvantages exist because of the natural characteristics of single-element ultrasonic
transducers. First, the parameters of the transducer focal region – the beamwidth and depth of
focus – are not adjustable. Second, the induced region of the pushing force is fixed unless there
is mechanical scanning of the single-element ultrasonic transducer, which is an inherently time-
consuming process. More recently, commercial 1D ultrasonic arrays with the capability to
generate a steerable beam, have been applied in OCE systems [70]. However, commercial 1D
arrays are designed for standard B-mode imaging, which requires wide bandwidth for high
resolution. As a consequence, the output energy of the generated ARF force is insufficient due

Fig. 1-9. Images of the posterior segment of the eye are acquired by the OCE system [69]. a: OCT intensity image that presents
the structure of the posterior segment of the eye. b ot d: Optical Doppler tomography raw data presented in time series that
shows the elastic wave (blue area) propagating across the tissue. The color bar presents the axial displacement.
In Figures b to d, ARF was applied to generate the elastic wave, whose propagation was presented as the displacement area
showing up in the different regions of the tissue at different times.

15

to the short pushing duration. In addition, array probes are not easily integrated with OCT
systems in practice due to the inability to align them confocally. Lastly, the 1D ultrasonic array
can only focus and steer the ARF beam along the azimuthal direction, resulting in significant
sidelobes in the elevational direction. In addition, acoustic microtapping (AμT), as a method
to generate ARF without using water as a medium because of the application of an air-coupled
ultrasound transducer, is another promising approach [40, 71, 72]. But the high transmission
voltage (up to 400Vpeak-to-peak) [71, 72] makes it difficult to design the array and system.
1.1.3 Stereolithography 3D printing technology
Three-dimensional printing technology has become one of the most rapidly developing fields
in recent years. Different from traditional formative manufacturing technologies such as
machining or forging, 3D printing technology implements a layer-by-layer deposition approach
instead of forming or casting and the design of the object can be generated by computer aided
design (CAD) software [73, 74]. With this approach, the manufacturing of complex structures
that required precision and control that are hard to be achieved by conventional fabrication
methods becomes possible [73, 75]. Additionally, 3D printing can support a wide range of
materials. From conventional thermoplastics, and resin to ceramics and metal [76], which
ensures its broad application scenarios. In short, the benefits of the 3D printing technology can
be summarized in the following aspects: 1. Direct translation from CAD design model to
physical objects; 2. The ability to generate highly customized designs without extra tooling or
cost; 3. Short lead time and product iteration time; 4. Supporting a wide range of materials for
fabrication [73]. This technology brings new opportunities and possibilities for its application in

16

both industry and academia. The implementation of 3D printing technology in industries will
significantly reduce the lead time, cost of human labor, and iteration time for the next
generation of products, which shows its potential of revolutionizing industries [74]. The short
lead time and direct translation from model to product also provide the opportunities for the
customers to make their specifications clearer than before to improve the efficiency for both
sides. Meanwhile, progress in 3D printing for high-quality and low-cost micro-scaler resolution
has already directly benefited a range of fields including biotechnology [77], optical devices
[78], and electronic devices [79, 80] to name just a few.
Different types of 3D printing technologies have been developed for varieties of applications
in the past few years. Based on ASTM standard F2792 [81], 3D printing technologies can be
categorized into 7 groups: binding jetting, directed energy deposition, material extrusion,
material jetting, powder bed fusion, sheet lamination, and vat photopolymerization. Each of
these types has its special application scenario. For example, binder jetting is suitable for

Fig. 1-10. Schemes of two typical stereolithography systems. Left: a bottom-up system using a scanning laser. Right: a top-down
setup implemented with digital light projection [89].

17

producing casting patterns and raw sintered products [74, 82, 83], while the 3D printing
technology based on material extrusion, the typical one is fused deposition modeling (FDM) 3D
printing, was good at multi-materials printing for plastics and food [84-86]. In our project, the
main 3D printing technology implemented is stereolithography (SLA) 3D printing, which
implements vat photopolymerization.
Stereolithography, a further type of 3D printing technology that is based on ultraviolet light
scanning and photosensitive resin solidification [Fig. 1-10] [87, 88], has shown excellent printing
resolution and demonstrated versatility in research and industrial work. Compared with other
3D printing technologies, stereolithography 3D printing presents its especially universal in
terms of the freedom to design architectures and the scales of the designs [89].
Stereolithography printers supporting various printing part sizes from sub-micron to decimeter
have been available and supported in different fields [90, 91]. Large 3D structures comprised of
nearly 10 million 150 μm pitch periodic cells, were presented by Angkur Jyoti Dipanka Shaikeea
et al [92]. To fabricate this design, a moving projection SLA system with the micrometer-scale
resolution was realized. 3D printing technology improvements are not limited to better
resolution. A novel STA technique which supported multi-material projection was reported by
Zhenpeng Xu et al [93]. This approach enabled dissolvable supports in complex architectures,
which makes the printing of overhanging, fly-like micro-architectures possible.
Stereolithography in particular has shown its great potential in the biomedical field. Benefits
from 3D printing technology's direct translation feature as well as supporting a variety of
complex printing structures and ability to manufacture internal structures precisely, SLA was

18

widely implemented in the fabrication of patient-specific models, surgery aids, and
biocompatible tissue engineering scaffolds [87, 89, 94, 95]. In this paper, we will use the many
advantages of SLA as mentioned above to assist in the fabrication of 2D ultrasound arrays
address the challenges they face and improve the overall fabrication process.
1.2 Motivations and Objectives
High-resolution SLA 3D printing technology enabled the fabrication of the interposer with
fine grid pitch [96] and made pitch-shifting interposers possible, which will become a novel
approach to address the challenges in 2D array fabrication. 3D printed grids filled with e-solder
have been proved by our previous work that can provide good conduction between the array
elements and the circuits (printed circuit board (PCB) or application-specific integrated circuit
(ASIC)) and work as the backing layer for the array and becomes a reliable method to build 2D
array [20, 96, 97]. Whereas SLA’s strengths in dealing with complex structures have not been
fully exploited. Re-examining the usage of interposers in a 2D array's manufacturing process
reveals that interposers are not simply used as components for connection as well as backing
layers, but provide a customizable component in the overall structure, which introduces a new
dimension between the transducer surface and the circuit surface. The interface between the
transducer and the circuits is exact the bottleneck of the 2D array fabrication which suffers
from the constraints of the high element density. PCB and ASIC act as the customizable
components in the array fabrication, nevertheless, the circuits do not support complex 3D
architecture and the lead time and cost are much longer and higher than the 3D printing part.
Additionally, PCB and ASIC cannot provide effective backing for the transducer. From the

19

perspective of image quality, embedded backing blocks between the circuits and elements are
desired. Based on the above points, further exploration of possible architectures of the
interposer and implementation them into 2D array fabrication for different types and
application scenarios will have many benefits for 2D array fabrication and be a powerful tool to
address the challenges in 2D array fabrication. Here, we present interposers with novel
architectures: pitch-shifting interposers [Fig. 1-11]. Different from interposers with straight
channels, pitch-shifting interposers were designed with an isometric scaling channel group,
which ends up in isometric scaling electrode matrixes on a different surface of the interposer.
Thus, the geometry of the pad matrix on the circuit is decoupled from the transducer aperture
and the constrain of the circuit area applied by the element size was released. This feature will
bring significant benefits to circuit design, especially for ASICs, which suffered from the trade-
offs between area, heat dissipation, and signal quality. Additionally, the implementation of the
interposers with reformatting channels brings another possibility to the array fabrication.
Traditionally, when ASICs or PCBs are designed and applied for a direct assembly process, the
circuits can only be used for certain designs. The aperture of the transducer is fixed, and a new

Fig. 1-11. Pitch-shifting interposer’s cross-section

20

array design necessitates redesigns of the IC design as well, which is extremely time-consuming
and costly. Nonetheless, 3D printed interposers are very affordable in terms of time and
material consumption. Regarding the interposers presented in the following chapters, once the
design is confirmed, the fabrication time is less than one week, including every step before the
final assembly. And the expenses are nearly negligible since the major consumable is the resin
and the consumption is minimal. Processes implemented with interposers are much more
affordable than IC design flow in many aspects.
In the following chapters, different interposers and 2D arrays design and fabrication
processes will be presented. With pitch-shifting interposer design, the restrictions on array
design have been significantly relaxed, thus giving us more options and the ability to explore
different array designs and special applications. The main objective of our works is to prove the
feasibility of the special interposers in array fabrication. Arrays with different operation
frequency ranges will be presented. Additionally, novel applications with the 2D array will be
implemented and the array’s performance will be examined.
1.3 Outline
This paper is organized into the following structure: In chapter 2, ring arrays’, which were
implemented in OCE, design, and fabrication process will be presented in detail. Including the
simulation of the acoustic field, the 3D models of the interposers, and important design steps.
In the last part of this chapter, a step-by-step introduction to the interposer and array
fabrication will be discussed. To test the quality and verify the function of the arrays. Each array
requires a pulse-echo test and hydrophone test. Mapping the pulse-echo results will provide

21

important information for the array quality such as yield and uniformity. Hydrophone
measurement results presented the intensity of the beam and details of the emitting field of
the array. Comparing the hydrophone results with the simulation results will be valuable for the
improvement of the next generation of array fabrication.
Chapter 3 will be the chapter presents the 2D ring array based OCE system’s setup and
experiment results. The combination of the OCT system and ultrasound control system will be
declared. The imaging results include phantom imaging ex-vivo and in-vivo results. The order in
which the results are presented follows the principle of gradual progression. From verifying the
accuracy of the system by phantom imaging to testing the experimental effects of actual
biological tissues in ex vivo, to finally testing the system performance in vivo experiments. This
new approach will be explained step by step.
Chapter 4 will introduce the design and fabrication of a high-frequency 2D array with a pitch-
shifting interposer. The idea of using this special interposer design to break the limitations in
circuit design from the array geometry will be realized in this chapter. Conventionally, a 2D
array designed at a high-frequency range requires the most advanced printed circuit fabrication
techniques or ASICs to meet its geometry requirements. In our work, a 2D array operated at 10
MHz was fabricated with a general PCB. This prototype becomes a strong testament to the
benefits of implementing interposers in high frequency. The details of its design, fabrication,
and performance testing will be included.

22

Chapter 5 is the conclusion part. Includes a summary of the presented work and a discussion
of the results. The possible future development of the technologies covered in this paper will
also be an important topic in this chapter.

23

Chapter 2 2D Ultrasonic Ring Array Fabrication with 3D printed
interposer

2.1 Array Design
As an initial attempt for 2D array-based AFR-OCT, a 2D ultrasonic array transducer with 3.5
MHz central frequency, 108 active elements, and 1.2 wavelengths (1.2 λ) pitch was designed for
this study, details of the array parameter are presented in Table 2-1. The aperture geometry
can be described as a 12 by 12 matrix with a 6 by 6 opening area [Fig. 2-1]. This array will be
referred to as G1 in the following. Since the array was designed for excitation, the requirements
of our array are different from arrays used for imaging.

Fig. 2-1. The geometry of G1 ring array: 12 by 12 matrix with 6 by 6 opening area at the central region of the transducer.

24

Single-element transducers have a narrow bandwidth, with which acoustic waves with long
pulse duration and relatively low frequency are generated and have been widely used for ARF
pulse excitation [67, 68]. In this study, a similar concept is applied to a fully electronically
controlled 2D array. To improve the performance for single-frequency operation forcing,
matching layers were not used at the front face of the array. Using the interposer as the
backing relaxes the required tolerance on the flatness of the surface of the printed circuit board
(PCB) or flexible printed circuit (FPC) for assembly. Normally, an interposer backing using
acoustically attenuating material will provide significant attenuation of the acoustic wave,
leading to reduced ringdown, and consequently a wider bandwidth at the expense of reduced
transmit sensitivity. This feature is advantageous for an imaging array, however not suitable for
this design in which our main goal is to provide efficient forcing at a single design frequency. To
match these requirements, we designed the interposer differently. To reduce the amount of
TABLE 2-1 Ring Array G1 Parameters
Parameter Value
Piezo Materials PZT-DL48 1-3 composite
Matching Layer N/A
Backing Interposer + E-solder
Center Frequency 3.5 MHz
Pitch 540 µm
Kerf 40 µm
Element size 500 µm × 500 µm
Number of Elements 108
Total Dimensions 6.48 mm × 6.48 mm
Central Opening Diameter 3.5 mm

25

backside attenuation, and thereby increase the resonance and output power of the elements,
the conducting interposer pillars were designed to be a smaller volume fraction of the
combined azimuthal and elevational pitch of the elements. The details of the interposer will be
discussed in the interposer section. The geometry of the matrix was designed as a trade-off of
the scanning window size and the element’s pitch. A large pitch will lead to strong sidelobes
which will significantly influence the stimulation, while the small scanning window will limit the
size of the field of view. As a result, the pitch was designed as 1.2 wavelengths with a 3.5 mm
scanning window.
Based on the data and experience gained from the G1. The parameter of the next generation
ring array has noticeable changes. For easy reference, this array will be referred to as G2. G2's
parameters (Table 2-2) are designed to address the following issues found in the experiments
TABLE 2-2 Ring Array G2 Parameters
Parameter Value
Piezo Materials PZT-DL48
Matching Layer N/A
Backing Interposer + E-solder
Center Frequency 1.5 MHz/2MHz
Pitch 1000 µm
Kerf 100 µm
Element size 900 µm × 900 µm
Number of Elements 256
Total Dimensions 20 mm × 20 mm
Central Opening Diameter 8 mm

26

with G1: 1. Narrow optical window. One typical application for OCE in ophthalmology is probing
the biomechanical properties of the corneal tissue, which usually has a diameter of around
10mm. A 3.5mm diameter window can barely cover over 10% of the whole tissue. 2. The
pressure generated by the acoustic beam was relatively weak. Based on the hydrophone
results, the pressure G1 ring array can provide with 40 V peak output voltage was 0.92MPa.
The most effective method to expand the size of the scanning window and the pressure of
the acoustic field is increasing the element number. The G2 ring array was designed with 256
elements. The aperture is a 20 by 20 matrix with an 8 mm scanning window. The elements at
the corners were not active so the geometry of the aperture will form into a ring shape, which
will help to keep the acoustic beam uniform during steering. The center frequency of the first
G2 ring array was set at 1.5MHz to match one wavelength pitch for the array elements. And for
the second G2 ring array, it was increased to 2MHz since the steering angle in the experiment
was usually less than 8°and ultrasound will be more efficient to generate ARF at a higher
frequency.
2.2 Field II Simulation
The ring element had been successfully fabricated and implemented in OCE already. Yet, the
G1’s design was limited by the element amount. Simulation results were necessary to study the
influence of the large pitch (1.2 wavelengths) and the opening window.
The geometry of the G1 ring array has been presented in the previous section [Fig. 2-1]. The
focal depth in the simulation was set at 15 mm according to the experiment setup. Since the

27

scanning area would not be larger than the size of the window, the maximum steering distance
from the central points of the XY plane was set to be 2mm.
The simulation results of G1’s emitting field in XY plane at 15mm depth with different
steering parameters were presented in Fig. 2-3. And the 1D plot of the cross-section at the focal
point for the first two simulations was shown in Fig. 2-4. According to the simulation results,
the -6dB intensity region at the focal point is 1.28mm. The -8dB sidelobes were found around
the main lobe. The sidelobes might be a potential issue for the experiment with high output
voltage since it could stimulate the tissue and lead to multiple shear waves on the surface.
Thus, the output voltage and maximum steer angle need to be limited in the experiments.
As described in the previous part, the G2 ring array was designed with new geometry [Fig. 2-
2]. The shape of the array is close to a ring and the element number was increased to 256 and

Fig. 2-2. The geometry of the G2 ring array: a 20 by 20 matrix with a round shape opening area. The extra elements at the
corners are not active, so the transducer’s geometry formed into a ring shape.

28

the opening window diameter was increased to 8 mm for the G2 ring array. Therefore, the

Fig. 2-4. 1D plot of the XY plane emitting field: a. Azimuth 0mm, elevation 0mm, depth 15mm; b: Azimuth -2mm, elevation
0mm, depth 15mm.

Fig. 2-3. XY plane of the emitting field of ring array G1 with different focus locations: a. Azimuth 0mm, elevation 0mm, depth
15mm; b: Azimuth -2mm, elevation 0mm, depth 15mm; b: Azimuth 1mm, elevation 0mm, depth 15mm; b: Azimuth 2mm,
elevation -2mm

29

maximum steered distance of the beam needs to reach 4mm to fully take advantage of the
enlarged scanning window. The simulation setup was applied for both 1.5MHz and 2MHz
designs. High operation frequency can significantly reduce the size of the -6dB region with the
risk of a strong sidelobe due to the mismatch between the pitch and the wavelength.
Comparing the C plane emitting field simulation results in Fig. 2-5, and 2-6 with Fig. 2-3 and
Fig. 2-4, we could see a clear change in the shape of the sidelobe. With the new geometry, the
sidelobes were reduced to -10dB and the acoustic field became more uniform. Reducing the

Fig. 2-5. XY plane of the emitting field and its 1D plot of 1.5MHz ring array G2 with different focus locations: a,b. XY plane with
the focal point at (Azimuth 0mm, elevation 0mm, depth 15mm) and (Azimuth -4mm, elevation 0mm, depth 15mm); c, d: 1D
plot of the focal point

30

operation frequency from 3.5MHz to 1.5MHz mainly affects the size of the focal point. The
diameter of the -6dB region increased to 1.48mm. Increasing the central frequency can
effectively solve this issue. In Fig.2-6, the diameter of the -6dB region is reduced to 0.9mm.
Nevertheless, sidelobe forms at the 6mm location as shown in Fig.2-6(d). Considering that the
maximum steering angle for the G2 ring array is 8°, the sidelobe in Fig.2-6(d) is the worst case
for the G2 ring array and since its intensity is -23dB, we believe it will not influence the

Fig. 2-6. XY plane of the emitting field and its 1D plot of 2MHz ring array G2 with different focus locations: a,b. XY plane with the
focal point at (Azimuth 0mm, elevation 0mm, depth 15mm) and (Azimuth -4mm, elevation 0mm, depth 15mm); c, d: 1D plot of
the

31

stimulation. However, further increasing the operation frequency will lead to an issue of a
strong sidelobe while steering the acoustic beam.

2.3 KLM model-based simulation
We used BioSono KLM 2.0 (Fremont, CA, USA), an online KML model-based simulator was
used to simulate the P/E results of each element based on the array design to confirm that a
long pulse duration will be achieved.

Fig. 2-7. Pulse-Echo impulse response simulation for G1 ring array

32

G1 ring array was implemented with a 1-3 composite. Based on simulation results [Fig 2-7],
its bandwidth will reach 34%. Without matching layers, the long pulse duration echo can be
seen in the pulse/echo results.
Compared with bulk material, 1-3 composite’s acoustic impedance was reduced. Thus, even
without any matching layer design, the G1 ring array’s bandwidth still reached 30%. To enhance
the long pulse duration characteristic of the transducer, bulk material was implemented for the
G2 ring array’s fabrication.
As shown in shown Fig. 2-8, the pulse duration was increased significantly in the pulse-echo
simulation results. And the bandwidth of the transducer was reduced to 19%. Based on the two

Fig. 2-8. Pulse-Echo impulse response simulation for G2 1.5MHz ring array

33

simulation results, the bulk materials are more applicable to our application. Nevertheless,
when we designed the 2MHz G2 ring array, we had to take consider the aspect of ratio the
1.5MHz G2 ring array, the aspect ratio of each element is close to 1. This design parameter will
influence the kt significantly already and reduce the efficiency of stimulation. In application, we
decided to increase the voltage to compensate for this issue to some degree. Yet, when the
central frequency was increased to 2MHz, the aspect ratio will reduce to 0.7. At this point, we
have to face this issue. Thus, for the 2MHz G2 ring array, we implemented 1-3 composite again.
As presented in Fig. 2-9, the duration of the pulse was reduced. The bandwidth of the
transducer reached 26%.

Fig. 2-9. Pulse-Echo impulse response simulation for G2 2MHz ring array

34

2.4 3D printed interposer design and filling
The interposer grid matrixes were designed with CAD tools. G1’s interposer is still a straight
channel interposer since the element density was still acceptable by the PCB process. The pitch
of the channels is 540 µm, which is matched with the element matrix [Fig. 2-10]. The
interposers designed for G2 were applied with a pitch-shifting architecture. This was our first
attempt at this structure for 2 main reasons: 1. The G2 requires 256 elements formed in a ring
shape matrix, which requires PCB technologies with fine trace and space. Using a pitch-shifting
interposer will reduce the requirement and can reduce the requirements for the
manufacturers’ capabilities and the cost; 2. G2 was designed as a standard model for the ring
array for stimulation. In future work, the design parameters of the aperture will need to be
changed according to different application scenarios. Combining the above two points, the
fabrication of G2 arrays will be a decent opportunity to test the pitch-shifting interposer design.
Presented in Fig. 2-11. f and g, the top (transducer side) and bottom (circuit side) of the

Fig. 2-10. 3D printed interposer. a: The interposer for the G1 Ring array. b: The interposer after e-solder filling and drilling.

35

interposer have different electrode matrixes since the pitch of the matrix on the top side is 1
mm and it is 1.5 mm for the bottom side. The interposers were fabricated by a ProJet 3500 HD
printer (3D Systems, Inc., Rock Hill, SC, USA) in XHD mode using VisiJet M3 Crystal acrylate
material. This printer can achieve 34 µm in X and Y dimensions and the resolution alone Z
direction (layer thickness) is 16 µm. In practice, the final size of the minimum structure usually
ended in 60 to 80 µm in X and Y dimensions. It is still sufficient for the design of the two
generations of the ring arrays and 2D array fabrication with denser elements matrix.
In the 2D array fabrication, the interposers act as the interconnection layer between the
material and the circuit. The most important parameter of the grid matrixes is the pitch
between the channels on the material side and circuit side. For the G1 ring array, the interposer
was designed with straight channels since the density of the elements was reduced with the
opening area [Fig. 2-10]. For the G2 ring array, we implemented the shift-pitch idea. The
element matrix on the PCB for G2 was designed with a 1.5 mm pitch to reduce the complexity
of both the PCB design and the assembly process. The channels in the interposer were designed
to be slop and therefore, on the material side of the interposer, a matrix with equal proportion
but 1 mm pitch was presented.
For each interposer, a frame with a slightly larger size than the interposer is designed [Fig. 2-
11. a]. This frame will work as a dam for the interposer filling. Usually, the height of this
structure needs to be 2 to 2.5 of the interposers.
The last step before the filling is the block detection for the interposer. Printing massive high
concentration channels is challenging work for 3D printers. The condition of the printers could

36

influence the quality and yield of the grid matrix significantly. Since the filling material is costly,
it is fatal to ensure the channels are all in good condition. The detection process is, however,
simple, and efficient: 1. Filling the interposer with water. Pushing the interposer into water is an
efficient method to achieve this and it is important to ensure all the channels are sealed with
water. 2. Blow the water out of channels by using compressed air from one side. 3. Checking if
there are still any channels filled with water. It is possible to clean the interposer with a wire if
some of the channels are blocked. However, this behavior may also damage the wall and lead
to a short between the channels. Giving up the blocked interposer will be the best way.
Once the interposers are confirmed with 100% yield, we can start the filling process [Fig. 2-
11. b-d]. At this step, the frame that was printed with the interposer will be useful. The frame
will be designed like a house with three walls and a “gate” [Fig. 2-11. a]. The frame includes the
“gate” that will be glued on glass with five-minute epoxy and the filling material will be pulled
into the frame. E-solder will involve many air bubbles during the mixing, and they will lead to

Fig. 2-11. Interposer filling, marking, and pins dicing. a: 3D printed frame. b, c: interposer filling and taking out process; d:
interposer for curing; e: Lapped interposer; f: Mark the surface; g: Pin matrix on the circuit side of the interposer after dicing

37

the disconnection in the channels. Due to the viscous nature of the e-solder, vacuuming does
not remove the air bubbles well because the cavity left after extracting the air bubbles still
exists and will not be filled with the e-solder. So, centrifuging will be the better choice here. As
described in the previous part, centrifuged e-solder with high attenuation was not necessary for
this array. Thus, the spinning speed for this step was set at 800 to 1000 rounds per minute and
the process lasted for only 5 minutes. In this situation, the e-solder paste did not experience
stratification but simply expel air bubbles. After centrifugation, we push the interposer into the
e-solder. With the help of the frame, the e-solder will be forced into the channels. When e-
solder flowed out through all the channels on the top side, the filling process is finished [Fig 2-
11. b]. The next step will be cutting off the “gate” and taking the filled interposer out. In
practice, we found that since the e-solder will shrink after curing, removing all the e-solder on
the surface of the interposer will lead to a concave surface on the channels, which will seriously
affect the quality of the finished product. Therefore, we avoided lifting the interposers when
removing them from the frame, but pushed them out from the side, and make sure that the top
and bottom surfaces are covered with excess resin during the curing. The filled interposer will
be cured for 24 hours at room temperature and then spend 2 hours in the oven at 45°C for
post-curing. Then we will lap the interposer with sandpaper and remove the extra e-solder on
the surface. Before assembly the piezo material to the interposer, two more steps need to be
done: 1. After the assembly, we will dice on the material along the walls between the channels.
Nevertheless, the grid matrix will have been covered by the material at that time. What is
more, the pitch of the grid matrix will be slightly different from what we designed in the 3D
model. So that the dicing may miss the walls and lead to shorts or open between the channels

38

without any mark. To avoid this, we diced on the surface before the assembly of the material.
In practice, the markers were not necessary for every cut. For example, for the G2 ring array’s
interposer, we left a mark every 5 pitches in each direction. The kerf will then be filled with dark
dyestuff, which makes it more recognizable; 2. Use thick blade dice on the circuit side of the
interposer to create separate pins. The thick kerf will help the interposer’s channels avoid being
shorted with each other. The details of this feature will be introduced in the following part.
Once the maker process and the pin creation are finished, the interposer is ready for the next
step.

2.5 Array assembly and fabrication
The array assembly contains two assembly processes. The assembly of the piezo material and
the interposer (material assembly) and the assembly of the acoustic stack and the PCB (acoustic
stack assembly). The key to a successful assembly is alignment. To achieve that, we designed a
customized fixture with a 2-axis stage combined with a rotation stage [Fig. 2-12]. And used a
USB microscope (Dino-Lite, Taiwan) to monitor the alignment process.

39

The material assembly was relatively easier compared with the acoustic stack assembly. The
interposer and the material were mounted on glass by wax and then fixed on the two ends of
the fixture with double-sided tape. The UCB microscope provided a close side view of the
channel grid and the edge of the material. To verify the alignment, we need to monitor all
available sides of the connecting part. The markers that were diced on the surface of the
interposer will be very helpful at this step. After confirming that the two parts were aligned, we
will spread the e-solder between the two parts and then lower the material on the interposer.
At this step, excess e-solder will be used for two reasons: 1. The extra e-solder will ensure a
good connection between the material and the interposer. Especially the two surfaces are not
flat to each other due to the unavoidable error involved by the fixture; 2. The excess e-solder

Fig. 2-12. Fixture for assembly. The base is built by combining the two stages and allows the adjustment of the location of the
object on the top of it. The sample holder can be taken out from the fixture, and usually used to hold the top part in the asse

40

will be squeezed out after lowering the material and will cover the periphery area of it, where
located with the GND channels. The extra e-solder will naturally become the GND electrode for
the array, which reduces the complexity of the array fabrication.
The fabrication of the 2D array happened between the two assembly steps. By the time the
material assembly is finished, all the channels are shorted together by an e-solder. To create
independent elements, we diced through the material and the e-solder layer and filled the kerf
with epoxy. To ensure the elements will not be shorted by the remaining e-solder, the kerf
needs to be deeper than the thickness of the material, usually, 100 μm will be a safe value. To
minimize the influence of the pitch variety. The dicing process for each direction should happen
twice: start from the center of the matrix and progress to the two sides. The marker on the
interposer will be used for alignment and calibration for this step. After the dicing of each
direction, the kerf needs to be filled with epoxy. The epoxy underneath each element may not
provide enough force to hold the element during the dicing. Dice two-direction continuously
may lead to the separation of the elements and the interposer.
Compared to the image array that will be presented in the following chapter, ring arrays
need one more step: creating the opening area. There are two different methods to achieve

Fig. 2-13. Array fabrication process. a: A close view of the material after first dicing; b: Acoustic stack after dicing in both
directions; c: Acoustic stack with drilled hole waited for sputtering; d: Sputtered acoustic stack with diced optical window.

41

this purpose. For the G1 ring array and the first G2 ring array, we followed the procedure
implemented for ring single transducers: using a drilling machine to remove the material.
Nevertheless, this method may lead to the depolarization and damage of the material.
Therefore, for the second G2 array, a dicing machine was applied to create the window.
According to the geometry of the element distribution, the four lines in the figure were chosen
[Fig. 2-13. d]. Then the array will be diced again on these four lines and then the window was
created. The kerf will later be filled with epoxy. The methods can be distinguished by the shape
of the opening area: the round shape in Fig. 2-13. c (drilling) and square shape in Fig. 2-13. d
(dicing).
The GND connection will be the last step for the array fabrication. For the G1 ring array, we
used a wire to connect the GND electrodes. This wire, unfortunately, will lead to much
inconvenience in the experiment setup. Thus, for G2, the GND connection method was
improved. GND channels were added to the interposer during the CAD modeling process.
During the material assembling, excess e-solder covered the channels connected to GND and
became the electrodes. After dicing and kerf filling, these GND electrodes will be disconnected
from the signal electrodes of the elements and has a continuous structure between the
electrode’s surface to the transducer surface. Thus, the GND connection can be finished by
simply sputtering gold on the surface of the array. This gold will connect the GND side of all
elements and the GND electrodes. Up to this point, the fabrication of the acoustic stack is
finished. And the last step is assembling the acoustic stack to the circuit.

42

The acoustic stack assembly process is more challenging than the composite assembling step.
Different from the material side that the composite is exposed and still available for further
process. The connection area between the interposer and the circuit is sealed. Any further
processing of this area will be very limited. It is not appliable to spread excess e-solder for the
connection, which will lead to short between the channels. To address this issue, we used a
“stamping” process to control the amount of e-solder that spreads on the pins. We first
prepared a “dam” on the surface of a glass. This “dam” is usually fabricated by a thin layer of
copper and the typical thickness is 50 μm. Later, an appropriate amount of e-solder will be
applied to the surface and then we will scrape across the surface of the dam with a hard and
flat object, such as a copper rod. With the help of the thin copper layer, a thin and flat layer of
e-soler will be left in the central area [Fig. 2-14. a]. After confirming the alignment between the
interposer and the circuit. The glass with the thin e-solder layer would be put in between the
circuit and the interposer. The interposer will be stamped on this thin e-solder layer first and
left. Then the glass with a thin e-solder layer would be removed and now the interposer can be
lowered on the circuit. The “stamping” process ensures that only an appropriate amount of e-

Fig. 2-14. The thin layer of e-solder before and after stamping a, b: e-solder layer before/after stamping; c: Close view of the
mark on the layer after stamping. A clear and uniform cross structure can be seen in each mark, and this indicates the stamping
process was successful.

43

solder will adhere to each pin of the interposer and form into e-solder bumps on the pins [Fig.
2-15. b]. Since the interposer and the circuit had been aligned, the pins on the interposer will
touch the pads on the circuit, and the e-solder paste will ensure the connection between them.
One important advantage of this process is that it reduces the requirement of the flatness of
the surfaces. The amount of the e-solder is determined by the thickness of the “dam” and
needs to be calibrated in practice if the size of the pin changes. If the layer is too thin, the yield
of the connection will be low since the insufficient e-solder will not cover all the gaps between
the pins and the pads on the circuit. And too much e-solder, however, will lead to short
between the channels. Because when the interposer was lowered on the circuit, the e-solder
will be squeezed out and excess e-solder may extend too much and short with near channels.
An efficient way to increase the yield of this assembly is by creating a thick kerf between the
pins on the interposer. A wide kerf will provide more space for the e-solder to spread and
increases the error-tolerant rate of the misalignment by reducing the size of the pins. For the

Fig. 2-15. The acoustic stack assembly process of Ring array G1. a: Overview of the setup; b: Upward view of the acoustic stack.
The silver bump on each pin is the e-solder bump. The photo presents the look of the bump that will lead to good conduction.

44

G2 ring array, the kerf between the pins reached 500 μm. In practice, we used two to
determine whether the result of the e-solder stamping process is eligible for assembly: 1.
Raising the interposer and viewing the pins. If the e-solder bumps on pins have a round surface
and are clear to be watched, it is ready for the next process [Fig 2-15]; 2. Watching the surface
of the thin e-solder layer on the glass. After an efficient contact with the pins, a mark will leave
on the surface. If the marks present a small cross in the center, it indicates that good e-solder
bumps formed on the pins [Fig. 2-14. b, c].
In both assemblies, the whole fixture remains locked during the curing [Fig. 2-16] and post-
curing (45 ℃, 2 hours). When the post-curing is finished, the whole fixture will be heated to
65 ℃, which is the melting point of the wax, and the array with the circuit will be removed from
the fixture and the array fabrication is finished [Fig. 2-17].

Fig. 2-16. Fixture with the array. The alignment has been finished and the acoustic stack has been assembled on the circuit.

45

2.6 Array test
The array performance was evaluated with pulse-echo (P/E) testing and hydrophone
measurement. For P/E testing, a quartz cube was used as the target and placed in a deionized
(DI) water-filled tank at the depth of 15 mm, which was also the focal distance in the following
tests and experiments. The array was securely attached to a controlled stage with micrometers

Fig. 2-17. Ring arrays. a: G1 ring array; b: G2 ring array (top: 2MHz G2, bottom: 1.5MHz G2).

Fig. 2-18. Typical pulse-echo response and spectrum of G1 ring array, operational frequency: 3.5MHz, bandwidth: 37%.

46

to adjust the position in X- and Y- dimensions. During the testing, one channel was turned on at
a time, with each of the elements interrogated in turn. The frequency of the monocycle pulse
was set at arrays’ operation frequencies. To ensure signal uniformity of the P/E results, array tilt
was corrected manually by monitoring the time-of-flight of each element and adjusting the
physical orientation of the array.
From the P/E results, the bandwidth and sensitivity of the elements will be yielded. The three
arrays’ typical elements’ pulse-echo results and their spectrum are presented in Fig. 2-18 to 2-
20. The P/E response waveform exhibited a long pulse duration, corresponding to a narrow -
6dB bandwidth. This is also evident in the fast Fourier transform (FFT) of the P/E waveform.

Fig. 2-19. Typical pulse-echo response and spectrum of G2 1.5MHz ring array, operational frequency: 1.5MHz, bandwidth: 20%.

47

Comparing the G1 ring array’s simulation results with P/E testing results [Fig. 2-18]. Increased
bandwidth will be noticed. There were two potential reasons to explain this change. The first is
the material performance deviates from the specification. The simulation results were based on
the parameters provided in the official specification and it is a common situation that the
products differ from the specification. The second reason could be the dicing process. The
alignment during the composite assembly process cannot be perfect, and the element may miss
off the grids. Thus, when we diced and create an individual element, extra active material could
be removed, and the acoustic impedance was further reduced.

Fig. 2-20. Typical pulse-echo response and spectrum of G2 2MHz ring array, operational frequency: 2.4MHz, bandwidth: 25%.

48

The tested spectrum of the 1.5MHz G2 ring array [Fig. 2-19] matched the simulation results
well. The 2MHz G2 ring array [Fig. 2-20], however, suffered from overlapping issues and has a
higher operating frequency than designed. Its bandwidth, though, did not mismatch too much
compared with the simulation results. Comparing the amplitude of the echo, the 2MHz G2 ring
array showed a stronger pulse than the 1.5MHz array, which should be the result of using
composite instead of bulk material.

49

Besides the typical element performance, the P/E results of the whole transducer array will

Fig. 2-21. Pulse-echo magnitude map of G1 ring array

Fig. 2-22. Pulse-echo magnitude map. a: G2, 1.5MHz array; b: G2 2MHz arrays. 6 centrosymmetric channels were open in
two G2 arrays. These were the results of the cable issue.

50

be used to generate the sensitivity maps, which represent the overall performance of the
arrays. The normalized echo magnitude will be plotted into a map according to the channel
map. As shown in Fig. 2-21 and 2-22, these maps present the yield of the array. The elements
with very low sensitivity are opened. We used these maps to determine if the arrays were
worthy of further testing or needed to be re-assembled. The presented array yields are all
above 90%. The main reason that leads to the open channels is the unevenness of the
electrodes on the interposers. The testing was finished with a 128-channel system. So for G2
ring arrays that have 256 elements, the testing will happen twice with the same probe. An
interesting phenomenon that can be observed in the 1.5MHz and 2MHz G2 ring arrays’
sensitivity map is there were 6 elements were not detected any signal in both arrays. And the
three elements on the left show a centrosymmetric distribution pattern with the three
elements on the right. Considering the truth that the pad distribution on the circuit followed
centrosymmetric distribution. We believe the issues that lead to the disconnection of these
elements came from the circuit part instead of the acoustic stack.

51

The arrays that passed the P/E test will then move to the hydrophone measurement. The
hydrophone measurement is carried out to evaluate the emitting field and verify the

Fig. 2-23. Top: illustration of the hydrophone measurement setup; Bottom: HGL-0085 hydrophone

52

customized Verasonics system control script. A hydrophone (HGL-0085, ONDA Corp.,
Sunnyvale, CA, USA) was securely attached to a 3-D stepper motor (SGSP33-200, OptoSigma
Corporation, Santa Ana, CA, USA), which was controlled by the PC to perform a full scan the
opening window area parallels to the array [Fig. 2-23]. A function generator (AFG 3252 C,
Tektronix, Beaverton, OR, USA) was used to synchronize the Verasonics system with the
digitizer card in the PC. The drive signal’s frequency was set at 3.5 MHz, 1.5MH, and 2.4 MHz
for G1, G2 1.5MHz, and G2 2MHz respectively, and the output voltage is 10 V Peak. The voltage
applied during the measurements was much lower than the voltage level during the
experiments to protect the hydrophone. During the testing, the focal depth was set at 15 mm,
which is a typical depth of the target in experiments.

Fig. 2-24. Hydrophone measurement results in XY plane of the ring array G1 (focal distance: 15 mm) with different steering
parameters: (a) original point; (b) y: 4°, x: 0°; (c) y: -4°, x: 0°; (d) y: 0°, x: -4°; e: y: 0°, x: 4°.

53

For each array, we measured its acoustic field with different steering parameters. As shown
in Fig. 2-24, 25, and 26. The focal point (red area) was moved in different directions. Comparing
the hydrophone results with the simulation results, the acoustic field of the G1 ring array
showed an obvious difference from the simulation. The reason for this difference could be the
yield and lower uniformity of the G1 ring array. 1.5 MHz G2 array’s acoustic field matched well
with the simulation. Besides, comparing the acoustic field of 1.5MHz G2 and 2MHz G2, the size
of the focal zone is reduced in the 2 MHz array’s results and its sidelobe is weaker, which makes
the 2 MHz array a better choice for experiments.

Fig. 2-25. Hydrophone measurement results in XY plane of the 1.5MHz ring array G2 (focal distance: 15 mm) with different
steering parameters: (a) original point; (b) y: 0°, x: 8°; (c) y: 4°, x: 0°

Fig. 2-26. Hydrophone measurement results in XY plane of the 2D array (focal distance: 15 mm) with different steering
parameters: (a) original point; (b) y: -8°, x: 0°; (360 μm by 360 μm per pixel)

54

Besides the shape of the acoustic field, another important performance specification
provided in hydrophone measurements is the pressure of the acoustic field at the focal point.
The OCE results’ quality highly relays on the pressure generated by the acoustic beam.
According to the hydrophone results, the G1, 1.5MHz G2, and 2MHz G2’s pressure at the focal
point (15mm focal depth and 40 V Peak output) are 0.92MPa, 1.9MPa, and 2MPa respectively.
From G1 to G2, the pressure provided by the acoustic was doubled.

2.7 Discussion
In this chapter, the details of the array design and fabrication process with the interposers
were introduced. And the arrays showed design-compliant performance in the testing. The
design and fabrication of arrays are one of the most important steps.
The design and fabrication of G1 and two G2 arrays were not closely linked. Thus, the step-
by-step improvements can be seen in the aperture design and array parameters. At the
beginning of the project, considering the many uncertainties and the limitations of the
hardware, G1's design was relatively conservative and had many immaturities. Its Limited
pressure and the strong sidelobes brought many limitations to the experiments. And that is why
we design and manufacture a 1.5MHz G2 ring array. Aimed at increasing the pressure and
reducing the sidelobe. The first G2 ring array was designed with more than doubled elements
compared to G1. The geometry of the element matrix was optimized to be closer to a ring
shape. In the simulation, the new aperture would significantly reduce the sidelobes and

55

reduced the size of the focal zone. After implementing the 1.5MHz array into experiments, we
realized that the higher operating frequency would bring a smaller focal zone and be more
efficient in stimulation. Additionally, 1.5MHz was right at the edge of the frequency range that
was supported by the Verasonics system. Whit these two reasons, a 2MHz G2 ring array came
into form.
Once a 2D stimulation array was finished, pulse/echo testing and hydrophone measurement
are the standard steps to evaluate their performance. Those critical probe characteristics, such
as bandwidth, central frequency, yield, acoustic field shape, and pressure at the focal point
yield from the testing results will be used to determine if this array needs to be
remanufactured. It is also important to compare the simulation results with experiment output
to evaluate the fabrication process. From, this we know that the dicing and assembly process
and material deviation all could lead to the mismatch between the array performance and the
simulation results.
With the standard 2D array fabrication process established based on the 3D printed
interposer, the 2D array becomes quite available. Nevertheless, some seemingly trivial steps
have a significant impact on the quality of the final product. Especially the filling step and
“stamping” step. In order to build conductive channels, an e-solder was used to fill the channels
in the printed structure. There are a few process details that must be emphasized here: 1.
mixing the two components of the e-solder will unavoidably cause air bubbles to be mixed into
the mixture, which will lead to open channels. Centrifugation will be an effective way to remove
these air bubbles. What is important is using different centrifugation parameters will let the e-

56

solder ends in different statuses. The operator must be aware of what type of e-solder is
desired before this step; 2. “stamping” process is a special approach developed with an
interposer for the 2D array fabrication. Compared with the conventional linear array attaching
method, the “stamping” process brings several benefits, such as reduced requirements for
flatness of the two surfaces, good matching with rigid circuits, etc. And the shortcomings are
also obvious, the whole process in the absence of mature equipment to assist the case, and the
final result is extremely dependent on the operator's proficiency and experience. It requires a
lot of experience to create a suitable e-solder layer for stamping and aligning and determine
whether the array is ready for assembly; 3. The pitch of the elements and the electrodes may
change due to the curing of the 3D printed part. Thus, the operator needs to be clear of the
suitable pitch of each side. It seems that the implementation of interposers leads to
considerable amounts of trouble rather than benefits when we consider these factors. Yet, the
truth is with the interposer, the requirement of creating a good connection between the
circuits and the transducer matrix is lowered. With the “stamping” process, the error tolerance
of the planarity has been increased. Additionally, the assembly process does not have to be a
one-time success. The acoustic stack can be removed without damaging the circuit or itself.
Another important point of the successful fabrication of these arrays, especially for G2, is
that this is the first time pitch-shifting interposers were implemented into array fabrication.
Even though the array’s designed frequency is not high, which reduces the difficulty. The usage
of the interposer demonstrates the feasibility of this architecture and shows the promising

57

possibility for future array design. As the next step, a high-frequency array, which is hard to
build with general PCBs, will be presented in chapter 4.

58

Chapter 3 2D array based OCE system Experiments results

3.1 Systems setup and imaging process
To successfully implement a 2D array-based ARF-OCE. The two systems need to be
synchronized. In practice, we used the OCT system as the master to trigger the Verasonics
system. The OCT system we applied is a customized 20 kHz SD-OCT system with a central
wavelength of 890 nm and bandwidth of 144 nm.
During the imaging process, the SD-OCT system sent out one trigger signal to the Verasonics
system for each lateral scanning location. At that location, a total of 400 A-lines (one M-mode
dataset) were acquired at 20 kHz. To establish the baseline or reference signal (i.e., the signal of
initial tissue position without the pushing force) for the axial displacement curve calculation,
the Verasonics system was excited 100 µs after the SD-OCT system started to acquire data. At
the Verasonics system side, a burst sinusoidal wave will be transmitted to the array when it was
triggered. The frequency of the sinusoidal wave was determined by the operating frequency of
the arrays according to their P/E results.
The Young’s modulus of the tissue can be found using the following equation:
E = 3ρc
2
(1)

59

where E is Young’s modulus, ρ is the density of the tissue, and c is the group velocity of the
mechanical wave. To obtain the biomechanical properties of the imaged phantom and tissues,
the group velocities of the shear wave were calculated by tracing the position of the wave
surface over time. The intensity of the shear wave, however, was not taken into consideration.
The raw data acquired by the scanning presented the location where the displacement
occurred in a 2D region (scanning direction versus axial direction) over time. To calculate the

Fig. 3-1. The fixture used to support the G2 array.

Fig. 3-2. Experiment setup for the G2 ring array.

60

group velocity of the shear wave, the raw data were first averaged to increase the signal-to-
noise ratio, and it was then re-sliced based on a custom algorithm. The processed results
represented the displacement along the scanning direction versus time, which indicates the
propagation of the shear wave at different depths. The group velocities of the shear wave can
be yielded by measuring the slope of the displacement curve.
In the experiments, the G1 ring array was fixed on a clamp. Whereas the same setup did not
work with the G2 ring array due to its size. To help the alignment of the G2 ring array, a 3D
printed fixture was used in the experiment setup [Fig. 3-1]. The final setup for G2 was
presented in Fig. 3-2.

3.2 Phantom Test and Ex-vivo Tissue Results
3.2.1 G1 ring array’s Phantom Results
Gelatin (Gelatin G8-500, Fisher Scientific International, Inc., Hampton, NH, USA) based tissue-
mimicking phantoms were designed and fabricated to verify the accuracy of the 2D array-based
OCE system. The homogenous phantoms comprised 7.5% gelatin and 0.5 % silicon carbide
powder (S5631, Sigma-Aldrich, St. Louis, MO, USA) as sound scatters. The stiffness of the
homogeneous phantom was measured using the gold standard of uniaxial mechanical testing

61

(5942, Instron, Norwood, MA, USA). The phantoms were cut into several small bulk units to fit
the device, and the average stiffness was 22.05 ± 0.73 kPa.
During phantom imaging, the output voltage of the Verasonics system was 35 V Peak with 8000
half-cycles. The size of the region of interest was mainly determined by the size of the optical
window in the acoustic 2D array. Fig. 3-3 (a) and (b) show OCT B-scan images with different
steering parameters. The dark area (marked by the vertical white line) is the position where the
traveling shear wave was initiated. Fig. 3-3(c) and (d) are the spatiotemporal maps. As can be
seen, by the comparison of Fig. 3-3(c) and (d), the blue area appears at different azimuthal
locations with a time delay, which indicates propagation of the mechanical wave at this depth.
The peaks in the images are the location where the excitation occurred. Comparing the
locations of the peaks in Fig. 3-3(c) and (d) to dark areas (white lines) in Fig. 3-3(a) and (b), we
find
that
the

Fig. 3-3. Phantom OCT B-scan results (a, b) and OCE results (c, d); [Steering parameters: (a) and (c) original point; (b)
and (d) elevation: 0, azimuth: 0.06 radian].

62

excitation occurred in a different location when the transmission parameters were changed.
This shows that the design functions properly, with the ARF beam being steered to stimulate a
specific area. Another important observation is that in all four images, there is only a single
excitation at the focal point. This indicates that the sidelobes which were found in the acoustic
hydrophone measurement [Fig. 2-24] did not influence the phantom images with the OCT
system. White lines in Fig. 3-3 (c) and (d) indicate the edge of the propagation trail of the
mechanical wave (blue area). To calculate the group velocity of the shear wave, dashed lines
were drawn along the blue area in Fig. 3-3 (c) and (d). As mentioned above, the group velocity
of the shear wave is given by the slope of the white lines in the spatiotemporal maps. The
group velocity was calculated to be 2.66 ± 0.08 m/s. Using Equation (1), the E of the phantom
was calculated to be 21.28 ± 1.3 kPa. Compared with the gold standard uniaxial mechanical test
result (described above), the error was less than 10%.
3.2.2 G2 ring array’s Phantom Results
Considering the size and the weight of the G2 array would make it difficult for alignment, we
designed a fixture to hold the transducer surface during the imaging process [Fig 3-4]. This
fixture ensured the array surface was parallel to the phantom’s surface and hold and support
the array avoiding damage to it during the setup. According to the hydrophone measurements,
both G2 arrays can provide higher pressure than the G1 array. Thus, we decide to compare the
performance between G2 and a concave ring element in G2’s phantom imaging experiments. A
concave ring element, which was implemented in our previous single element ARF-OCE, was

63

used for comparison and has a 4.5 MHz central frequency, 20mm outer diameter, and 8mm
inner diameter. To drive this element, a burst consisting of 2700 cycles of sine wave was
generated by a function generator and amplified by a power amplifier to 45V peak. The G2 array
was driven by the Verasonics system with a signal containing 2500 half-cycles sin wave at
2.3MHz with 40V peak output voltage.

Fig. 3-4. G2 phantom imaging experiment setup.

64

The images were further processed to present the edge of the shear wave clearer. The group
velocity was measured to be 8.21 m/s in Fig. 3-5 (a) and 8.03 in Fig. 3-5 (b). And Young’s
modulus of the two measurements is 202.0 kPa and 194.1 kPa respectively. The difference
between these two measurements is within 5%. For the G1 ring array, a rigid target with
200kPa Young’s modulus was beyond its capabilities. With the G2 ring array, a comparable OCE
result was acquired. Though the limitation is still clear. Comparing two images, the one
acquired with a ring array showed low contrast than the single element’s results, which are the
results of different pressure. With the same output voltage level, single elements still show
higher pressure than a 2D array. 2 main reasons lead to this difference. First is the impedance
mismatch. In a 2D array’s geometry, each element’s size has to be n by n, where n is usually a

Fig. 3-5. Single element (a) and G2 array (b) phantom results comparison.

65

value from half a wavelength to twice a wavelength. The limited size will lead to severe
impedance mismatch between the probe and the system, which in turn reduces the efficiency
of power transfer. Another single element’s advantage is it has a large active area. Although the
two transducers have a similar inner and outer radius, the structure actually gives a more active
area to the single element. Besides, the thickness of the propagation lines in the G2 array’s
result is thicker than the single element’s result. The possible reason led to this phenomenon is
that the single element with a higher operating frequency provides a smaller focal zone
compared to the 2D array. Thus, the shear wave showed a cleaner propagation trail.
With a larger aperture and more elements compared to the G1 ring array, G2 presented a
comparable performance in the aspect of providing acoustic radiation force compared to a
single element. What single elements cannot do is control the beam in real-time without
involving mechanical scanning. Yet, the limitation of the 2D array is still existing and influences
the imaging results, which can be clearly seen in the images. This is, to some degree, a trade-off
between the performance and the capabilities of beam controlling.
3.2.2 G1 ring array’s Ex-vivo tissue results
After verifying the script’s function and accuracy of the system, the next step was testing the
performance with real tissues. Different from phantoms, the ocular tissues, such as the cornea,
do not contain any scatters. In phantom imaging, the scatters help the target received the ARF
from the acoustic beam. Without them, the shear wave will become very weak and influence
the final image quality. As a consequence, the voltage used for tissue imaging was usually
higher than what was applied in phantom imaging. Unfortunately, for the 2D system, the

66

available voltage was quite limited compared to single elements, which were driven by power
amplifiers. Thus, even though we confirm the function of the 2D array based OCE system with
phantom, uncertainty remains as to whether compliant images can be produced in the tissues.
The ex-vivo tissue we used in the experiment was a healthy rabbit excised eyeball collected
from a local slaughterhouse (Sierra Medical Science, Inc., Whittier, CA, USA) within 12 hours of
death. Phosphate-buffered saline solution was used to preserve its freshness and as a medium
for ultrasound coupling. All experiments were performed at room temperature. The system
setup was similar to phantom imaging. Expect drive signal was changed with a lower cycle
number (2000 half-cycles) and higher output voltage (40 V Peak).
The blue vertical lines in Fig. 3-6 (a) and (b) marked the stimulated point on the tissue. And
similar to the process for phantom images, we used white dashed lines marking the edge of the
propagation trail of the shear wave. Based on the dashed lines, the group velocity of the shear
wave was calculated to be 1.38 ± 0.1 m/s, and Young’s modulus of the tissue was yielded with
Eq. (1) and its 5.72 ± 0.42 kPa. Note that Eq.(1) can be applied only if the target is a

Fig. 3-6. G1’s Ex vivo OCE imaging of a rabbit eye: (a) and (b) spatiotemporal displacement map; (c-e) shear wave
propagation map.

67

homogeneous bulk material [98]. The Young’s modulus is listed here for reference only. The
group velocity of the tissue matched with previously reported measurements [99]. shear wave
propagation vs. elapsed time is shown in Fig. 3-5 (c) to (e), where T = 0 is the time point when
the initial trigger from the OCT system was first sent to the Verasonics system. The blue color
showed up in the images indicating the displacement led by a shear wave. It was first generated
in Fig 3-5 (c) and its propagation was presented in the following two figures.
With the images, the effectiveness of ring array design in tissue stimulation experiments was
confirmed. The contrast of these images was not ideal. A sharp and clear edge of the
propagation trail is desired for accurate measurement of the group velocity of the shear wave.
And this limitation mainly came from the pressure that the G1 ring array could provide.
Compared with phantom images, the contrast issue became worse since the cornea lack of
the scatter to interact with the acoustic wave. Thus, an array with higher pressure than
G1 was necessary to yield more accurate biomechanical properties of the tissue. And
this is the exact motivation for us to design and fabricate the second generation 2D ring
array for the 2D array based ARF-OCE system.
3.2.2 G2 ring array’s Ex-vivo tissue results
For the same reason, the G2 array was implemented in tissue imaging after the experiments
with the phantoms. The same rabbit eyeballs collected from the local slaughterhouse (Sierra

68

Medical Science, Inc., Whittier, CA, USA) within 12 hours of death were used as the targets in
this experiment.
The setup of the ex-vivo experiment was similar to the phantom experiments. Expect the
fixture was added with a holder to hold the eyeballs. With a larger scanning window compared
to G1, G2’s scanning window covers the major area of the corneal tissue. Due to the curved
structure of the cornea, the tissue would be out of the focal zone of the OCT system at the end
of the scanning. This was the reason for the dark area on the right side of the three images in
Fig.
3-7.

Fig. 3-7. Spatiotemporal displacement map of Ex-vivo tissues with ring array G2: (a) and (b) were acquired at different
locations; (c) Same location as (a) with less cycle number of the driven signal.

69

The G2 array was driven by the Verasonics system with a signal containing 2500 half-cycles
sin wave at 2.3MHz with 42V peak output voltage, and the beam was steered by 2° in Fig. 3-7 (b).
Fig.3-7 (c) was acquired at the same location as (a), except the cycle number was reduced to

Fig. 3-8. Fixture and system setup test. Top: The fixture to hold the rubber sheet for the pop-out process. Bottom: Pre-
alignment for the system before setup the animal.

70

400 half-cycles. With the higher pressure, the G2 ring array can generate a clear shear wave
without a large cycle number and acquire biomechanical properties with less power delivered.
The group velocity of the shear wave was measured to be 1.40 ± 0.4 m/s and Young’s modulus
of the cornea was calculated to be 5.88 ± 0.39 kPa, which is very similar to the previous results.
Considering the cornea’s properties, this Young’s modulus was provided for reference.
Benefiting from the higher pressure, the images yielded with the G2 ring array presented
higher contrast compared with G1’s ex-vivo results. The edge of the shear wave’s propagation
trail was sharp and clean. And with the wider scanning window, G2’s results presented more
area of the tissue compared to G1.
3.3 G2 ring array’s In-vivo tissue results
Considering the pressure level provided by G1, the experiments stopped at ex-vivo. After
comparing the experiment results with the single element and ex-vivo tissues, the G2 ring array
showed great potential for the in-vivo study. A New Zealand white rabbit was used as an
experimental subject. The main difficulty for the experiment setup was that system need to
work with a clean and appropriate medium, which is transparent for laser, to couple the array
with the target. Ultrasound gel was commonly used for transducer coupling and Its
characteristics seem to fit the application scenario here. Unfortunately, the unavoidable air
bubbles and the rough surface made it not suitable for this setup. Eventually, physiological
saline solution was used as the medium for this experiment. Yet, how to place the in-vivo tissue
and the array in the solution became the new problem. To make this setup work, the pop-out
process became necessary. The fixture presented in Fig. 3-8 top was used to hold the rubber

71

sheet for the pop-out process. Before anesthetizing the animal, the system will be set up first to
adjust each part and roughly align the array and the camera [Fig. 3-8 bottom]. After the animals
were anesthetized and the eyes were properly treated. The eyeball would be popped out of the
eye socket. A rubber sheet with a suitable size hole in the middle would be used to fix the
eyeball. The sheet was mounted on the fixture in Fig. 3-8 and the solution would be poured into
the pit formed with the rubber sheet. In this way, the array was coupled with the tissue by the
solution. The finished setup is presented in Fig. 3-9. Top.

Fig. 3-9. In-vivo experiment setup and OCE result. Top: Experiment setup. The rabbit’s eye was popped out and fixed with a
rubber sheet. Bottom: Spatiotemporal displacement map of the cornea. The short yellow string represented the edge of the
shear wave transmission.

72

The pop-out process will significantly increase the intraocular pressure, which will make the
tissue stiffener than in relaxed conditions, thus it required higher pressure to successfully
generate a shear wave on the target. Nevertheless, the power limitation of the system makes
the higher voltage become not applicable. The consequence is the in-vivo image [Fig. 3-9.
Bottom] has a lower contrast compared with ex-vivo results. The driven signal applied in this
experiment was 2500 half-cycles sin wave at 2.3MHz with 44V peak. The Young’s modulus of this
in-vivo tissue was measured to be 34.2 kPa. The increase in Young’s modulus was the result of
the pop-out process.
3.4 Discussion
In this chapter, the experiment setup and imaging results of the co-registered 2D ultrasonic
array-based OCE system were demonstrated. OCE images and reasonable biomechanical
properties of different targets were yielded with the G1 ring array. To the best of our
knowledge, this is the first reported result for a system capable of generating a controllable
spatiotemporal pushing force using a fully programmable 2D ultrasonic array. In response to
the multiple problems and limitations revealed in G1’s experiments, we developed and
implemented G2 in similar experiments. By comparing the results with single elements and G1’s
images, the shear wave generated by the G2 ring array was comparable to that of a single
element at a similar drive voltage when stimulating a hard phantom, while the ex-vivo image
quality was significantly improved compared to the G1. Additionally, for the first time, in-vivo
OCE results were yielded with the G2 ring array.

73

Phantom imaging is a standard step to evaluate the array and the system setup. Mixed with
scatters, the phantoms have a good response to the acoustic radiation force, which can well
reflect the quality of the shear wave generated by the acoustic field corresponding to the
transducers. In addition, the phantoms’ Young’s modulus can be easily measured by applying
uniaxial mechanical testing and their structure can be designed to be suitable for generating
shear waves. All these properties made phantom imaging a very important step to test and
caliber the system’s performance. The G1 ring array’s phantom imaging experiment had
another important goal that was examining the system setup. Before this step, all the acoustic
tests and measurements were finished with the Verasoncis system alone. How to combine and
set up the trigger signal properly for the ultrasound system and the OCT system was still
unknown at this stage. The results yielded with the G1 array with phantoms were compared
with the gold standard measured results. With the acceptable error between the two methods,
the system setup was confirmed with the G1 array. Regarding the G2 ring array, the system
setup has been finished and tested with G1, the phantom imaging experiments were focused
on the G2’s performance testing. The overall constraint of the system setup with G1 was the
insufficient strength of the acoustic field. Especially compared with the single elements. Thus,
we decided to compare the array output with the single element. Considering the impedance
mismatch and active area, it would be very difficult to achieve identical performance between
an array and a single element. Nevertheless, the images were still comparable. The target was a
rigid phantom, and both methods acquired a clear shear wave trail. Though the array image
showed a lower contrast, capable of producing clear shear waves on the rigid phantom showed
the potential of the setup with the G2 ring array.

74

Ex-vivo studies were the preparation step for in-vivo studies and the last step for the G1.
Lacking the scatter, tissues usually show a weak response to the ARF compared with phantoms.
The imaging parameters used for phantom imaging were not able to be directly used for tissue
imaging. Thus, ex-vivo tissue experiments were necessary to adjust the parameters and confirm
the feasibility of in-vivo experiments. Rabbit eyes’ cornea tissues were selected as the imaging
objects. The influence of the target properties can be clearly seen by comparing G1’s phantom
results and ex-vivo tissue results. The contrast of the shear wave trail become worse, and the
propagation distance of the shear wave was shorter in the ex-vivo tissue images, both were
indicating the lack of the intensity of the acoustic beam generated by G1. In conclusion, we
believe that the G1 ring array was not suitable for in-vivo study, which provided higher
requirements on the pressure of the stimulation beam, and a new generation with a larger
aperture and more elements was necessary for further experiments. Thus, the G2 was designed
and fabricated. The situation improved significantly in the G2 experiments. As presented in the
hydrophone measurement results, the pressure and the focal point size were improved. These
improvements in image quality are reflected in better contrast and signal-to-noise ratio
compared to G1’s results. Additionally, the G2 provided a larger scanning window that could
cover the whole cornea.
As described above, due to the limited pressure, the G1 was not implemented in in-vivo
studies. The in-vivo study was only successful with the G2. Nevertheless, the image quality was
still very limited. To have the tissues coupled with ultrasound transducers, we have to use the
pop-out process with a special fixture. The fixture made the setup trickier and the pop-out

75

process led to much higher intraocular pressure compared with the ex-vivo experiments. As a
consequence, the tissue becomes more rigid than in relaxed conditions and with a similar
output voltage, the intensity of the shared wave is reduced significantly. Potential methods to
improve the results are 1. Using a system that can provide higher driven voltage. The
Verasoncis system applied in this experiment is a 256-channel system without an external
power supply. To stimulate the tissue, the driven signal usually needs to last for a few hundred
microseconds. The system, however, was designed for the conventional ultrasound imaging
process. The output signal would easily reach the maximum power limitation of the system
with this duration and the reachable maximum voltage level was about half of the maximum
voltage supported by the system. Therefore, there is still a lot of room for improvement in
raising the output voltage, which will significantly increase the pressure of the acoustic field. 2.
Optimizing the experiment setup to avoid the pop-out process. Well-preserved ex-vivo tissues
should have very similar properties to the in-vivo tissues. The main reason that the in-vivo
tissue in experiments show high Young’s modulus was the abnormal intraocular pressure rising
led by the pop-out process. And the pop-out process was a reluctant solution to couple the
transducer with the eyeball. For an ultrasound transducer with few megahertz operating
frequencies, a couple of material is necessary. Even though we discuss the reason for using
physiological saline solution, this unwanted intraocular pressure rising forced us to seek a
solution of avoiding the pop-out process. A potential method is using ultrasound gel. The
problems of applying gel in this setup have been declared before. Nevertheless, there are still
several methods that may solve these problems. By degassing and centrifuging the gel multiple
times and spreading it carefully on the tissue, it is possible to reduce the air bubble

76

concentration to an acceptable level. And by applying optical glass on the surface of the gel, a
clean and regular interface can be created. The main limitation of these approaches is that they
will lead to significant changes in the optical path. The updates and adjustments to the OCT
system are necessary.

77

Chapter 4 10MHz 2D array based on the pitch-shifting interposer

4.1 2D array and pitch-shifting interposer design
Pitch-shifting interposer has been implemented in G2’s design and fabrication. As a
prototype, the application of this architecture mainly focuses on feasibility Verification. It is an
advantage, however, was not presented fully. In this chapter, the design and fabrication of a
high frequency 2D array based on a pitch-shifting interposer will be introduced. Benefiting from
TABLE 4-1 10MHz 2D array parameters
Parameter Value
Piezo Materials Soft PZT, 1-3 composite
k t 0.6
First Matching Layer 2-3 μm silver epoxy
1
ST
ML Sound Speed 1961 m/s
1
ST
ML Acoustic Impedance 7.84 MRayl
1
ST
ML Thickness 47 μm
Second Matching layer ABS
2
ND
ML Sound Speed 1850
2
ND
ML Acoustic Impedance 2.2 MRayl
2
ND
ML Thickness 44 μm
Backing E-solder
Center Frequency 10 MHz
Pitch 300 µm
Kerf 60 µm
Number of Elements 256
Yield 88%

78

pitch-shifting architecture, the requirements of the circuit design were reduced significantly
and the whole design is more available and affordable compared to the conventional
fabrication process.
This array was designed as a high-frequency 2D imaging array. A 1-3 composite made with
soft PZT material (DL-53, DeL Piezo Specialties, LLC, West Palm Beach, FL, USA) and epoxy as
kerf filler (EPO-TEK 301, Epoxy Technology, Billerica, MA, USA) were used as the active material
of this array and two matching layers (2-3 μm silver epoxy and ABS) were implemented. The
details of the array design can be found in Table. 4-1. Considering the size of the element, single
crystal materials could be a better choice for this application. Yet, the process of building 2D
arrays would be too risky for them. For this very first attempt, we aimed at a more stable
output.

Fig. 4-1. Conceptual drawings for 16 by 16 2D arrays and interposers. Top: array matrix and element’s structure; Bottom: Cross-
section of the pitch-shifting interposer

79

The array aperture had a 16 by 16 matrix [Fig. 4-1. Top] with a center frequency of 10 MHz as
a trade-off between the field of view and image quality, the element pitch of the array was
designed to be 2  (300 μm). Larger arrays with the same field of view at  pitch could be
implemented by increasing the number of imaging channels and this would improve image
performance by reducing side lobes. The interposer was designed with an isometric
magnification structure and is mainly formed with parallel channels where the pitch increased
from 635 μm (at the PCB side) to 300 μm (at the transducer side) [Fig. 4-1. Bottom]. The top
and bottom were straight channel regions each with a 16  16 matrix with matching pitch to
the pad matrix on the PCB and the aperture. The thickness of the walls between the channels
was 235 μm and 60 μm for the PCB side matrix and the transducer side matrix, respectively.

Fig. 4-2. Pulse-Echo impulse response simulation for 10MHz 2D array.

80

The wall thickness to pitch ratio was optimized differently for the PCB vs. the transducer matrix.
On the transducer side, a thinner wall was preferred for a more active area of the interposer
backing for better acoustic absorption. The thicker walls on the circuit side reduce the risk of
short circuits between channels during the assembly process for a similar reason described in
the G2’s interposer design part. In the center region, the channels were scaled isometrically and
the distance between the channels changed from 635 μm to 300 μm.
BioSono KLM 2.0 (Fremont, CA, USA) was implemented to evaluate the acoustic performance
of the array with the design parameters. The pulse-echo simulation result was presented in Fig.
4-2. With the two matching layers design, the bandwidth of the elements would reach 78%.
We left relatively long lengths for the straight regions. This design was intended to provide
more margin for the following steps and the straight channels connected to the 10 MHz
elements acted as the backing layer. A thicker backing provides improved attenuation of
acoustic energy on the back of the array which is especially important in imaging for reducing
ringdown and thereby improving axial resolution.
4.2 Circuit design
High-density interconnect printed circuit boards (HDI-PCB) were used to route the elements
to the system cable [Fig. 4-3]. These industry-standard substrates have a 2+N+2 architecture, 3
mils (75 m) trace and space and are capable of being fabricated by general manufacturers.
Compared with conventional PCBs, HDI-PCBs are smaller in size and provide different via types
instead of through holes only, which are advantageous in reducing routing complexity and

81

increasing the density of the pad matrix. The PCB used for array connection has a 16  16 pad
matrix with a 635 μm pad pitch in the central area, which will be used for array assembly in the
following process.
4.3 Interposer fabrication
The interposer for the high-frequency array [Fig. 4-4] placed high demands on the print
resolution. The minimum printing structure (channel wall at array side) in the interposer was
designed as 60 μm. And the channels’ length was designed as 10 mm to efficiently reduce the
ring down, which makes the printing task more challenging for the printer. In practice, the
minimum resolution of the 3D printers can only be achieved under certain conditions. Thus, to
successfully print this interposer, a printer providing finer resolution than the minimum
structure size is necessary. For this project, we implemented a DLP 3D printer (Kudo, Dublin,
CA, USA) with a minimum resolution of 15 μm to fabricate this design. After printing, the
interposer was immersed in alcohol and placed in an ultrasonic cleaner for 15 to 20 minutes.
Due to the porous structure of the interposer, this step is crucial to clean out any remaining
resin in the channels and avoid the channels being blocked. The interposers are not fully cured

Fig. 4-3. High-density interconnect printed circuit board for the 2D array.

82

after printing and therefore, after cleaning they were further cured in an oven at 46°C for one
hour. Every interposer was checked again to ensure clean channel yield after curing.
The filling process is similar to the ring array’s interposer. The Dam for containing the silver
epoxy filling material was fabricated to match the outer dimension of the interposers and was
glued to a supporting glass substrate for easy handling. The dams were filled with silver epoxy
and the interposers were carefully lowered into the dam-encapsulated epoxy. An important
consideration in this regard is that the dams need to fit snugly with the interposer outer
dimension with minimal gaps between the two. If the gap is not large enough, the interposers
cannot be lowered due to friction against the walls, while too large a gap would cause leakage
in the filling process and lead to the failure of the interposer fabrication due to insufficient
filling of the channels and wasted material.
The fabricated dam/glass substrate structures were filled with conducting silver epoxy paste
(E-Solder 3022, Von Roll Isola, New Haven, CT, USA) and centrifuged at 1000 rpm for 5 minutes.
Centrifuging is important for consolidating the silver epoxy particles to increase the acoustic
attenuation of the backing material. After this process, the paste surface was flat, and any air

Fig. 4-4. Interposer for the 10MHz 2D imaging array a: interposers before e-solder filling; b, c: Interposer filled with e-
solder; b: Transducer side; c: Circuit side.

83

bubbles created during the mixing process were removed. Uniformity of the backing material
with no air bubbles is important to maintain high-quality acoustic attenuation without any
significant echoes back to the transducer composite. The interposers were pushed into the
dam-confined silver epoxy slowly and with evenly applied high pressure to ensure that the
paste fills every channel. Yield in the fabrication of 2D arrays is multiplicative, and every step
must have a target yield of 100% such that the final yield of the overall process is high.
Therefore, we optimized the fabrication process of the interposers to achieve the highest yield
possible and further validated the electrical functioning of every channel in the completed
structures before continuing the overall transducer fabrication process.
After the interposer was pushed to the glass, the dam was cut, and the interposer was
released. Filled interposers were left in a dry box for 24 hours for curing. Extra silver paste on
the surface of each active site of the interposers was desired since the paste shrinks during
curing and can lead to dents and holes in the channels. Filling with a given interposer can only
be done one time as an attempt to rework an incompletely filled channel invariably leads to an
air gap that isolates the previously filled paste and new paste in the channels, and this reduces
the yield of the conductive channels significantly. After curing, the silver paste in the channels
hardens and forms a conductive pathway. The cured interposers were then lapped at the top
and bottom surface to remove the extra e-solder and create a flat surface on each side [Fig. 4-4.
b, c]. Surface co-planarity is critical for both the subsequent transducer fabrication steps as well
as the assembly to the electronic substrate. We therefore incorporated surface measurements
into the post-fabrication testing criteria with a goal of co-planarity below +/-10 m. The last

84

step in the fabrication process was dicing on the surface to create isolated assembly pins with
air gaps between them [Fig. 4-5. a]. We further diced the surface to create fiducial marks which
are critical for precise alignment [Fig. 4-6. a] for the subsequent steps in the overall acoustic
module fabrication process.
4.4 Array fabrication
The fabrication process for the acoustic stack was similar to the ring arrays except that the
imaging array had the first matching layer attached before assembly and did not have an
opening area. The first matching layer was a mixture of Insulcast 501 (Insulcast,
Montgomeryville, PA) and 2-3 μm silver powder. The mixture was cast on the composite while
placed in a dam and centrifuged. After curing, it was lapped to the design thickness of 47 μm.

Fig. 4-5. Schematic of Acoustic Stack Fabrication, the most left and right pillars are connected to GND: a: Creating
electrodes on the circuit side of the interposer; b: Assembling Piezo material (with first matching layer for imaging
array); c: Creating separated elements and filling kerf; d: Sputtering gold; e: Attaching second matching layer; f:
Finished acoustic stack.

85

Once the piezo material was ready, it was bonded on the transducer side of the interposer
using a thin layer of E-solder [Fig. 4-5. b]. Interposers with attached materials are shown in [Fig.
4-6. a]. The E-solder paste was applied to the interposer surface first, then the material was
aligned and pressed down on the paste. The extra paste squeezed out and surrounded the
piezo material. This is by design, as the channels connected to the ground were designed to
surround the central channels. The extra paste connects with those channels and is used for
ground connection in subsequent fabrication steps. The assembly was left in a dry box
overnight for curing. Once cured, the piezo material was diced to create independent elements.
The marks created in previous steps were used for the alignment of the dicing saw to the
interposer grid [Fig. 4-4. c]. This process was applied for both directions of the matrix and EPO-
TEK 301 was used as kerf filler to support the elements [Fig. 4-6. b]. It is critical to avoid dicing
along the edge of the material, as in this case a short circuit would be formed between the
elements on the edges and ground [Fig. 4-6. b]. We, therefore poured additional kerf filler in
these kerfs to build a smooth bridge from the element surface to the E-solder that covers the
ground. A matching layer detaching issue may happen at this step. As shown in [Fig 4. e], the

Fig. 4-6. The array fabrication process of the imaging array. a: piezo material assembled; b: dicing along with the marks on the
interposer to create individual elements; c: array assembled on HDI-PCB. The columns exhibiting gold color in e were suffering
from a matching layer detaching issue, which was the consequence of loss of connection at the interface between the matching
layer and the gold layer or the interface between the gold layer and the piezo material. The second matching layer was
assembled in c.

86

columns show gold color in locations where the matching layer was delaminated. There could
be more columns affected since not every detached matching layer is removed during the
dicing process. This issue has the potential to influence the bandwidth and sensitivity of the
array, which is observed in the testing results. After the curing process of EPO-TEK 301, the
array was sputtered with a Cr/Au (50/100 nm) electrode on the surface of the matrix [Fig. 4-5.
d]. By doing so, the top electrodes of all the elements and the extra E-solder connected to the
ground were connected. The “bridge” mentioned earlier is necessary to make this connection
possible and its shape directly influences the integrity of this connection. The above-described
process steps fully describe the fabrication of the pushing array. For the imaging array, an
additional step is added to the fabrication process which is a lamination of the second matching
layer [Fig. 4-5. e].
Finished acoustic stacks were next assembled to the PCBs. The PCB was designed at USC
using the Altera CAD tool and fabricated by an external vendor (MKT Electronic Co., Ltd.,
Guangdong, China). The “stamping” process introduced in our previous work [96] was used for
assembly. The circuit side of the interposer was first stamped on a thin layer of E-solder and
then the acoustic stack was aligned and assembled on the matrix area of the HDI-PCB. After
curing, the array fabrication process was completed [Fig. 4-6. c].
4.5 Performance testing and imaging results

87

A similar process was undertaken to acquire the imaging array’s P/E results. The sensitivity
map was created based on the data and the resulting measured overall yield was 88% [Fig. 4-7.

Fig. 4-7. Array performance maps. a. Sensitivity map of the imaging array; b. Bandwidth map of the imaging array.

Fig. 4-8. Typical pulse-echo response and spectrum of G1 ring array, operational frequency: 11.6MHz, bandwidth: 44%.

88

a]. Most of the disconnection happened at the edges, which was usually a consequence of the
overlapping of the electrodes on the edges. Another factor that influenced the assembly
process was the area of the array. During the assembly process, the array surface was waxed on
a holder. This array’s aperture size was less than ¼ of the stimulation arrays, which made it
more difficult to mount the array flatly on the holder. The central left area presented lower
bandwidth compared with other elements [Fig. 4-7. b], which was due to matching layer
detachment. A typical element’s pulse-echo was presented in Fig. 4-8. The average bandwidth
of the working elements is 41%, which shows a significant mismatch between the simulation
results. Besides the detaching issues and overlapping. The main reason for the mismatch was
that the element after dicing cannot be treated as 1-3 composite anymore. Due to the

Fig. 4-9. Pulse-Echo impulse response simulation for 10MHz 2D array with bulk material.

89

technology limitations, the 1-3 composite we bought has a 180 μm pillar size. After the
alignment and dicing, the array elements were formed with one pillar instead of several pillars
with epoxy kerf. In this condition, the element should be treated as composite but bulk
material. Fig. 4-9 presented the updated simulation result. The material was changed to bulk
material and the bandwidth dropped to 60%.
The array was next used to image the wire targets. A target with five micro-scale steel wires
was used as the phantom to evaluate the imaging array. The wires were arranged in a
trapezoidal pattern with a vertical spacing of 1.5 mm and horizontal spacing of 0.5 mm [Fig. 4-
10]. The frames of the target provided a 6 mm by 6 mm imaging window that matches the
geometry of the imaging array. Coherent plane-wave compounding was implemented as the
beamforming technique. Each imaging cycle generated 21 angled plane waves in the X-Z plane
and Y-Z plane respectively. Images in the X-Z plane, Y-Z plane, and XY plane and a volume image
were reconstructed. Image reconstruction was processed by the Verasonics system in real-time.
Imaging views for XZ, YZ, and XY planes and a volume image were acquired in real-time [Fig. 4-

Fig. 4-10. Wire target with 5 micro-scale stainless steel strings. The strings were mounted on the stepwise structure of the
frame tightly.

90

11]. XZ and YZ planes correspond to the cross-section that crossed the y-axis and x-axis
respectively, and the XY plane was located at Z = 60 wavelengths. In the displayed volume

Fig. 4-11. Images of the wires target. a-d were acquired with the wires parallel to the x-direction, and e-h were acquired after
the target was rotated by 45 degrees. a, e: XY plane; b, f: XZ plane; c, g: YZ plane; d, h: volume images.

91

image [Fig. 4-11. d], the matrix formed by a group of black points at the top represents the
array aperture and provides aid for recognizing the position of the target relative to the
transducer. Two sets of images were recorded [Fig. 4-11 a-d and Fig 4-11 e-h]. The target was
rotated by 45° in the second image set, which can be seen by the orientation of the wires in XY
images [Fig. 4-11. a. e] and 3D models [Fig. 4-11. d. h] in volume images.

4.6 Discussion
In this chapter, the design, fabrication, and imaging performance of the interposer based
10MHz 2D array was introduced. The array presented in this work was an example to show the
promising potential of implementing pitch-shifting interposer architectures and 3D printing
technology in ultrasound array fabrication. Pitch-shifting interposers were designed following
the same idea in our previous work: filling the channels with a conductive epoxy to create
conducting channels that act as a backing layer [20, 96, 97]. In addition, with the introduction of
a shape transformation along the Z-axis, the interposers are no longer simply for connection
and backing, and now also play a role in shaping the array geometry and pitch requirements.
The pitch-shifting architecture is a straightforward modification of the original idea of
interposers: by applying isometric scaling of the pitch between channels and the diameter of
the channels, this design methodology allows us to adapt the fixed pad matrix on the circuit
board to any number of different pitch array matrices. The transformation that is inherent in
pitch-shifting also introduces new and potentially far-reaching topological possibilities provided
by interposers. Specifically, the deformation achieved in such interposers is not limited to pitch

92

scaling alone and can be further used for reshaping the geometry of the aperture. For example,
previously described spiral sparse-array geometries [35] could be implemented using this
technique. In general, any re-mapping of the transmission and receive element location that
was previously constrained by the ASIC pad array definition, can now be realized as long as the
geometry of the circuit electrodes provides a one-to-one mapping with the target element
matrix, and the structure is supported by the 3D printer used to create the interposer.
Fabrication of filled interposers as described here is a (sometimes messy) hands-on process
that is limited to prototyping and therefore not readily amenable to volume production. For the
quick-turn realization of novel element mapping geometries, this prototyping capability is
invaluable to rapidly test and validate design concepts. In the future, this process could be
adapted to volume production using injection-molded interposer die parts and industrial epoxy
filling extrusion machines.
In such a volume process it would be critical to identify important yield issues including
bubbles formed in the epoxy fill which reduce the effectiveness of the acoustic backing. In
addition, the yield of electrical connections is critically dependent on issues such as pinholes
formed in the walls of the channels which can potentially lead to shorting of neighboring
elements. Finally, coplanarity of the top and bottom surface of the interposers is a particularly
important yield criterion because it has significant effects on the assembly of both the top side
transducer array as well as the bottom side electronic substrate connections. All these issues
must be taken into consideration if a viable production process is to be implemented.

93

In addition, the active material (soft PZT) applied in this project was not suitable for 2D array
fabrication. The limitation of the dicing machine led to the “bulk material” issue, which was
described in the array performance section, and significantly influenced the performance of the
array. Considering the array size, the single crystal material produced by the etching process
would be promising for this application.
We expect that future exciting improvements in 3D printing technology will bring more
possibilities to interposer design. With new printers that support multiple material types,
acoustic backing fill and lapping steps could also be simplified, which will further improve the
yield and quality of the interposer.
In summary, with the use of advanced 3D printing technologies, we have implemented novel
3D printed pitch-shifting interposers which enable an adaptive 2D array fabrication process in
which the circuit pad topology does not limit the geometry of the transducer matrix. With the
refinement of the fabrication process and improvements in 3D printing technology, interposers
with continuous deformation of the channel matrix will bring new possibilities for array design
and solve circuit design challenges for dense 2D array implementation at low cost and with
quick-turn prototyping capabilities.

94

Chapter 5 Summary and Future Work

5.1 Summary and discussion
In the projects presented in previous sections, several array designs implemented with
interposer were fabricated, tested, and implemented in various applications.
With the ring array, the first 2D array based OCE system was built. A relatively small scanning
field with the requirement for high intensity and controllable acoustic beam in ARF-OCE
technology made this application well matched with the ring array design. Promising data was
acquired with the first-generation ring array design. Images of phantom and ex-vivo tissues
demonstrated the accuracy and effectiveness of the system and indicated the direction of
improvement as well. Thus, the second-generation ring arrays were designed and built based
on the issues that were exposed with the first attempt. With a larger aperture, more elements,
and a wider scanning window compared to the G1 ring array, the images were enhanced with
better contrast and a wider ranger. And the data from in-vivo tissue was acquired. Due to the
pop-out process, the in-vivo tissue had a higher Young’s modulus compared to ex-vivo tissues
and required stronger ARF to generate the shear wave. Nevertheless, limited by the system, the
output voltage level was not able to reach a similar level of single elements driven signal. This
issue would be one of the important future works for this project.
Without interposers’ help, the fabrication of the 2D ring arrays would be exponentially more
difficult. Higher error tolerance with more aperture geometry options made the ring array’s

95

fabrication become a smooth, efficient, and easy iterative process, which provides a solid
foundation for the success of the experiments. It is also an important test for the feasibility of
pitch-shifting interposer in 2D array fabrication, which became the cornerstone of the following
project.
The ring arrays applied for ARF-OCE were designed with relatively low central frequency.
Nevertheless, what makes the pitch-shifting interposer architecture promising for 2D array
fabrication is the possibilities it provided for high-frequency array design. Limited by the strict
requirements on element scale, the 2D arrays with high operation frequencies rely on printed
circuit technologies with high resolution or ASIC [Fig. 5-1], which could be costly and time-
consuming. Additionally, the circuits, either PCB or ASIC, with the determined structure can be
implemented for a certain array design only. New designs indicate new circuit design, which is
expensive and time-consuming. Introducing pitch-shifting interposers into array design and
fabrication can address the two challenges effectively. Benefiting from the rapid development
of 3D printing technologies, interposers with delicate and complex structures can be produced
quickly at an affordable cost. Compared to circuit design, the cost of redesign and trial and
error is significantly reduced. As a prototype, a 10 MHz fully sampled 2D array was fabricated
and tested. The array was tested and implemented into volumetric imaging for the wire target.
The pitch-shifting structure made it possible to build this array with a PCB with standard

96

capabilities. In the future, by adjusting the ratio of size variations, the same circuit can be used
for a variety of circuit designs.

Fig. 5-2. Interposer for 20MHz 1.75D array [100]. a: Before filling; b: After filling.

Fig. 5-1. Interconnection ASIC.

97

5.2 Future Work
In recent years, 3D printers that support objects with minimum structure from 10μm to
submicron were gradually available and affordable for research groups. Besides the 10MHz
array’s interposer presented in this paper, a prototype pitch-shifting interposer for a 20MHz
1.75D array was introduced in our previous work [100] [Fig. 5-2]. This interposer was printed by
Printer P130 (Boston Micro Fabrication, Maynard, Massachusetts, USA). On the transducer side,
it provided an 8 by 16 matrix and 50 μm by 130 μm element size with a 20 μm wall. The filling
and further processing for this tiny interposer were challenging, yet not impossible. With these
two examples, the capabilities of the 3D printer in interposers fabrication have been proven. In
the future, with the help of an advanced 3D printer, we could push on these technologies to
fabricate 2D arrays with even higher operating frequency. Additionally, the development of
novel 3D printers that support multi-material printing and printing materials also brings us
possibilities in optimizing the interposer’s fabrication process, e.g., building the interposer in
one shot with the multi-material printers. By avoiding the filling process and applying new
material as a backing layer to improve the yield and performance of the array.

98

With more choice of materials for printing, the design of the interposer itself could be
changed as well. The original design of the interposer was filling the 3D printed structure with
the conductive epoxy to form channels that can connect the transducer with the circuit and
serve as the backing layer at the same time. This design idea places high demands on the choice
of the filling materials: a conducting material with suitable acoustic impedance and high
acoustic attenuation. It would be very difficult to find a replacement for an e-solder that is
workable for 3D printing technology. In addition, 3D printing technologies with high resolution
was capable of printing small and solid structure or holes with a small diameter. Regarding the
interposer, however, the minimum structure is always the thin wall between the channels. This
is a great challenge for 3D printing technology, and the result is that the minimum wall
thickness is often several times larger than the print resolution, which could be a waste of the
capability of the 3D printer.

Fig. 5-3. The interposer with separated functional parts.

99

Thus, another promising interposer design is dividing the function of providing the current
path and providing backing into separate structures of the interposer. The material with
suitable acoustic impedance and attenuation could be more available. And using small holes fill
with an e-solder or inserted with thin wires to provide connection will fully take advantage of
the high resolution [Fig. 5-3]. The main challenges of this design would come from two aspects:

Fig. 5-4. 3D super-resolution imaging based on mechanical scanning 1D array. Top: system setup; Bottom: 3D images of the
rabbit eye’s posterior pole close to the optical nerve head. a: Power Doppler Imaging with microbubble; b: super-resolution
microvessel imaging; c, d: The corresponding vessel distribution superposed on the B-mode imaging. [101]

100

1. Finding the material that is compliant with acoustic characteristics and compatible with 3D
printing technology; 2: Building the conductive channel with a diameter of tens of microns.
Besides the next generation high frequency 2D array fabrication. Applications with the 2D
array will be another important part of our future work. In our previous work [101], 3D super-
resolution images were acquired with an imaging system based on a mechanical scanning 1D
array [Fig. 5-4]. This work showed promising results yet there were several limitations in this
system setup: 1. The vessels in the eyeballs are distributed in the plane parallel to the scanning
direction. Thus, the resolution of the vessels distributed along the scanning direction was
limited by the scanning step; 2. Mechanical scanning could lead to the motion of tissue during
the imaging process, which would significantly influence the accuracy of the structural
information in the images; 3. One volume image was generated by acquiring and combining
multiple 2D super-resolution images, which would lead to a very long acquisition time. The
microbubble concentration would change significantly during the imaging process and causes
inconsistent image quality from different slices. In the worst case, multiple injections were
necessary to maintain the microbubble concentration at a suitable level for data acquisition.
Implementing a 2D array for 3D super-solution imaging will help us overcome these
constraints. A 2D array based imaging system does not need mechanical scanning to acquire
data from 3D space. Thus, the imaging process could avoid the influence of mechanical motion.
Meanwhile, the acquisition time of a volume image will be comparable to the acquisition time
of a 2D image with a linear array and all the slices would be acquired within one acquisition,
which could avoid the change in image quality and multiple injections.

101

In addition, using a 2D array to achieve 3D elastography will be another interesting topic. In
conventional ultrasound elastography, biomechanical properties of the target can only be
measured within one direction with 1D arrays. By expending the sampling point in the 2D
plane, the 2D array could track the propagation of the shear wave in a plane. It will contribute
to the accuracy of the measurement and better response to the anisotropy of the target.
Regarding the 2D array based OCE system. The work till now has proven the function and
accuracy of the setup. In the future, implementing an OCT system with better resolution and
faster acquisition speed and ultrasound systems that support higher output voltage will further
improve the image quality and will make real-time 3D OCE possible. Applying new aperture
geometries and experimenting with different operating frequencies will potentially improve
image results and allow more application scenarios for the system. It is also important to
develop new approaches to set up the system to avoid the pop-out process for in-vivo studies.
Additionally, 2D stimulation arrays have shown promising abilities in the beam controlling for
stimulation experiments. Exploring its application for other stimulation applications could open
entirely new research directions.

102

BIBLIOGRAPHY
[1] J. A. Jensen, “Medical ultrasound imaging,” Progress in biophysics and molecular biology, vol. 93,
no. 1-3, pp. 153-165, 2007.
[2] D. G. Wildes, and L. S. Smith, "Advanced ultrasound probes for medical imaging." pp. 801-808.
[3] K. K. Shung, Diagnostic ultrasound: imaging and blood flow measurements: CRC press, 2015.
[4] C. M. Otto, The practice of clinical echocardiography: Elsevier Health Sciences, 2007.
[5] M. H. Mozaffari, and W.-S. Lee, “Freehand 3-D ultrasound imaging: a systematic review,”
Ultrasound in medicine & biology, vol. 43, no. 10, pp. 2099-2124, 2017.
[6] G. M. Treece, A. H. Gee, R. W. Prager et al., “High-definition freehand 3-D ultrasound,”
Ultrasound in medicine & biology, vol. 29, no. 4, pp. 529-546, 2003.
[7] G. M. Treece, J. E. Lindop, A. H. Gee et al., “Freehand ultrasound elastography with a 3-D
probe,” Ultrasound in medicine & biology, vol. 34, no. 3, pp. 463-474, 2008.
[8] P. Yan, J. C. Cheeseborough III, and K. C. Chao, “Automatic shape-based level set segmentation
for needle tracking in 3-D TRUS-guided prostate brachytherapy,” Ultrasound in medicine &
biology, vol. 38, no. 9, pp. 1626-1636, 2012.
[9] R. Ohbuchi, and H. Fuchs, "Incremental volume rendereing algorithm for interactive 3D
ultrasound imaging." pp. 486-500.
[10] R. Ohbuchi, D. Chen, and H. Fuchs, "Incremental volume reconstruction and rendering for 3-D
ultrasound imaging." pp. 312-323.
[11] J. Hung, R. Lang, F. Flachskampf et al., “3D echocardiography: a review of the current status and
future directions,” Journal of the American Society of Echocardiography, vol. 20, no. 3, pp. 213-
233, 2007.
[12] J. Yen, and R. Wodnicki, "Orthogonal Bowtie-Shaped 2D array for Real-time 3D Imaging." pp. 1-
4.
[13] I. E. Timor-Tritsch, and L. D. Platt, “Three-dimensional ultrasound experience in obstetrics,”
Current Opinion in Obstetrics and Gynecology, vol. 14, no. 6, pp. 569-575, 2002.
[14] C. Errico, J. Pierre, S. Pezet et al., “Ultrafast ultrasound localization microscopy for deep super-
resolution vascular imaging,” Nature, vol. 527, no. 7579, pp. 499-502, 2015.
[15] J. Zhu, E. M. Rowland, S. Harput et al., “3D super-resolution US imaging of rabbit lymph node
vasculature in vivo by using microbubbles,” Radiology, vol. 291, no. 3, pp. 642-650, 2019.

103

[16] J. A. Jensen, M. L. Ommen, S. H. Øygard et al., “Three-dimensional super-resolution imaging
using a row–column array,” IEEE transactions on ultrasonics, ferroelectrics, and frequency
control, vol. 67, no. 3, pp. 538-546, 2019.
[17] S. Harput, K. Christensen-Jeffries, A. Ramalli et al., “3-D super-resolution ultrasound (SR-US)
imaging with a 2-D sparse array,” arXiv preprint arXiv:1902.01608, 2019.
[18] J. O. Fiering, P. Hultman, W. Lee et al., “High-density flexible interconnect for two-dimensional
ultrasound arrays,” IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol.
47, no. 3, pp. 764-770, 2000.
[19] C. Chen, Z. Chen, D. Bera et al., “A Front-End ASIC With Receive Sub-array Beamforming
Integrated With a $32\times 32$ PZT Matrix Transducer for 3-D Transesophageal
Echocardiography,” IEEE Journal of Solid-State Circuits, vol. 52, no. 4, pp. 994-1006, 2017.
[20] R. Wodnicki, H. Kang, D. Li et al., "Tiled large element 1.75 D aperture with dual array modules
by adjacent integration of PIN-PMN-PT transducers and custom high voltage switching ASICs."
pp. 1955-1958.
[21] P. Guo, F. Fool, E. Noothout et al., "A 1.2 mW/channel 100µm-Pitch-Matched Transceiver ASIC
with Boxcar-Integration-Based RX Micro-Beamformer for High-Resolution 3D Ultrasound
Imaging." pp. 496-498.
[22] D. Wildes, W. Lee, B. Haider et al., “4-D ICE: A 2-D array transducer with integrated ASIC in a 10-
Fr catheter for real-time 3-D intracardiac echocardiography,” IEEE transactions on ultrasonics,
ferroelectrics, and frequency control, vol. 63, no. 12, pp. 2159-2173, 2016.
[23] J. T. Yen, J. P. Steinberg, and S. W. Smith, “Sparse 2-D array design for real time rectilinear
volumetric imaging,” IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol.
47, no. 1, pp. 93-110, 2000.
[24] H. Yoon, and T.-K. Song, “Sparse rectangular and spiral array designs for 3D medical ultrasound
imaging,” Sensors, vol. 20, no. 1, pp. 173, 2020.
[25] O. Martínez-Graullera, C. J. Martín, G. Godoy et al., “2D array design based on Fermat spiral for
ultrasound imaging,” Ultrasonics, vol. 50, no. 2, pp. 280-289, 2010.
[26] E. Roux, F. Varray, L. Petrusca et al., “Experimental 3-D ultrasound imaging with 2-D sparse
arrays using focused and diverging waves,” Scientific reports, vol. 8, no. 1, pp. 1-12, 2018.
[27] C. H. Seo, and J. T. Yen, “A 256 x 256 2-D array transducer with row-column addressing for 3-D
rectilinear imaging,” IEEE transactions on ultrasonics, ferroelectrics, and frequency control, vol.
56, no. 4, pp. 837-847, 2009.
[28] M. F. Rasmussen, and J. A. Jensen, "3D ultrasound imaging performance of a row-column
addressed 2D array transducer: a simulation study." p. 86750C.

104

[29] K. Latham, C. Samson, J. Woodacre et al., “A 30 MHz, 3D Imaging, Forward Looking Miniature
Endoscope based on a 128-element Relaxor Array,” IEEE Transactions on Ultrasonics,
Ferroelectrics, and Frequency Control, 2020.
[30] S. W. Smith, R. Davidson, C. D. Emery et al., "2-D array transducers for medical ultrasound at
Duke University: 1996." pp. 5-11.
[31] B. Savord, and R. Solomon, "Fully sampled matrix transducer for real time 3D ultrasonic
imaging." pp. 945-953.
[32] J. W. Choe, Ö. Oralkan, and P. T. Khuri-Yakub, "Design optimization for a 2-D sparse transducer
array for 3-D ultrasound imaging." pp. 1928-1931.
[33] B. Diarra, M. Robini, P. Tortoli et al., “Design of optimal 2-D nongrid sparse arrays for medical
ultrasound,” IEEE Transactions on Biomedical Engineering, vol. 60, no. 11, pp. 3093-3102, 2013.
[34] S. Harput, K. Christensen-Jeffries, A. Ramalli et al., “3-d super-resolution ultrasound imaging
with a 2-d sparse array,” IEEE transactions on ultrasonics, ferroelectrics, and frequency control,
vol. 67, no. 2, pp. 269-277, 2019.
[35] A. Ramalli, E. Boni, C. Giangrossi et al., “Real-time 3-D spectral doppler analysis with a sparse
spiral array,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 68, no.
5, pp. 1742-1751, 2021.
[36] E. Roux, A. Ramalli, P. Tortoli et al., “2-D ultrasound sparse arrays multidepth radiation
optimization using simulated annealing and spiral-array inspired energy functions,” IEEE
transactions on ultrasonics, ferroelectrics, and frequency control, vol. 63, no. 12, pp. 2138-2149,
2016.
[37] M. Flesch, M. Pernot, J. Provost et al., “4D in vivo ultrafast ultrasound imaging using a row-
column addressed matrix and coherently-compounded orthogonal plane waves,” Physics in
Medicine & Biology, vol. 62, no. 11, pp. 4571, 2017.
[38] K. Latham, C. Ceroici, C. A. Samson et al., “Simultaneous azimuth and Fresnel elevation
compounding: A fast 3-D imaging technique for crossed-electrode arrays,” IEEE transactions on
ultrasonics, ferroelectrics, and frequency control, vol. 65, no. 9, pp. 1657-1668, 2018.
[39] J. T. Yen, and E. Powis, "Boundary Array Transducer and Beamforming for Low-Cost Real-time 3D
Imaging." pp. 1-4.
[40] M. A. Kirby, I. Pelivanov, S. Song et al., “Optical coherence elastography in ophthalmology,”
Journal of biomedical optics, vol. 22, no. 12, pp. 121720, 2017.
[41] X. Qian, T. Ma, C.-C. Shih et al., “Ultrasonic microelastography to assess biomechanical
properties of the cornea,” IEEE Transactions on Biomedical Engineering, vol. 66, no. 3, pp. 647-
655, 2018.

105

[42] D. Dascalescu, C. Corbu, M. Constantin et al., “Correlations between Corneal Biomechanics and
Glaucoma Severity in Patients with Primary Open Angle Glaucoma,” Maedica, vol. 10, no. 4, pp.
331, 2015.
[43] K. R. Pillunat, C. Hermann, E. Spoerl et al., “Analyzing biomechanical parameters of the cornea
with glaucoma severity in open-angle glaucoma,” Graefe's Archive for Clinical and Experimental
Ophthalmology, vol. 254, no. 7, pp. 1345-1351, 2016.
[44] B. R. Costin, G. P. Fleming, P. A. Weber et al., “Corneal biomechanical properties affect
Goldmann applanation tonometry in primary open-angle glaucoma,” Journal of glaucoma, vol.
23, no. 2, pp. 69-74, 2014.
[45] S. R. Flaxman, R. R. Bourne, S. Resnikoff et al., “Global causes of blindness and distance vision
impairment 1990–2020: a systematic review and meta-analysis,” The Lancet Global Health, vol.
5, no. 12, pp. e1221-e1234, 2017.
[46] P. D. Jaycock, L. Lobo, J. Ibrahim et al., “Interferometric technique to measure biomechanical
changes in the cornea induced by refractive surgery,” Journal of cataract & refractive surgery,
vol. 31, no. 1, pp. 175-184, 2005.
[47] P. S. Binder, R. L. Lindstrom, R. D. Stulting et al., “Keratoconus and corneal ectasia after LASIK,”
Journal of refractive surgery, vol. 21, no. 6, pp. 749-752, 2005.
[48] T. Seiler, K. Koufala, and G. Richter, “Iatrogenic keratectasia after laser in situ keratomileusis,”
Journal of refractive surgery, vol. 14, no. 3, pp. 312-317, 1998.
[49] W. W. Haw, and E. E. Manche, “Iatrogenic keratectasia after a deep primary keratotomy during
laser in situ keratomileusis,” American journal of ophthalmology, vol. 132, no. 6, pp. 920-921,
2001.
[50] I. I. Bussel, G. Wollstein, and J. S. Schuman, “OCT for glaucoma diagnosis, screening and
detection of glaucoma progression,” British Journal of Ophthalmology, vol. 98, no. Suppl 2, pp.
ii15-ii19, 2014.
[51] Z. Gatzioufas, and B. Seitz, “Determination of corneal biomechanical properties in vivo: a
review,” Materials Science and Technology, vol. 31, no. 2, pp. 188-196, 2015.
[52] M. Singh, J. Li, S. Vantipalli et al., “Noncontact elastic wave imaging optical coherence
elastography for evaluating changes in corneal elasticity due to crosslinking,” IEEE Journal of
Selected Topics in Quantum Electronics, vol. 22, no. 3, pp. 266-276, 2015.
[53] D. A. Luce, “Determining in vivo biomechanical properties of the cornea with an ocular response
analyzer,” Journal of Cataract & Refractive Surgery, vol. 31, no. 1, pp. 156-162, 2005.
[54] B. F. Valbon, R. Ambrósio, B. M. Fontes et al., “Ocular biomechanical metrics by CorVis ST in
healthy Brazilian patients,” Journal of refractive surgery, vol. 30, no. 7, pp. 468-473, 2014.

106

[55] J. Ophir, I. Cespedes, H. Ponnekanti et al., “Elastography: a quantitative method for imaging the
elasticity of biological tissues,” Ultrasonic imaging, vol. 13, no. 2, pp. 111-134, 1991.
[56] A. L. McKnight, J. L. Kugel, P. J. Rossman et al., “MR elastography of breast cancer: preliminary
results,” American journal of roentgenology, vol. 178, no. 6, pp. 1411-1417, 2002.
[57] D. L. Cochlin, R. Ganatra, and D. Griffiths, “Elastography in the detection of prostatic cancer,”
Clinical radiology, vol. 57, no. 11, pp. 1014-1020, 2002.
[58] M. Yoneda, K. Suzuki, S. Kato et al., “Nonalcoholic fatty liver disease: US-based acoustic
radiation force impulse elastography,” Radiology, vol. 256, no. 2, pp. 640-647, 2010.
[59] X. Qian, T. Ma, M. Yu et al., “Multi-functional ultrasonic micro-elastography imaging system,”
Scientific reports, vol. 7, no. 1, pp. 1-11, 2017.
[60] C.-C. Shih, X. Qian, T. Ma et al., “Quantitative assessment of thin-layer tissue viscoelastic
properties using ultrasonic micro-elastography with Lamb wave model,” IEEE transactions on
medical imaging, vol. 37, no. 8, pp. 1887-1898, 2018.
[61] R. M. Lerner, K. J. Parker, J. Holen et al., "Sono-elasticity: Medical elasticity images derived from
ultrasound signals in mechanically vibrated targets," Acoustical imaging, pp. 317-327: Springer,
1988.
[62] J. Ophir, S. K. Alam, B. S. Garra et al., “Elastography: imaging the elastic properties of soft tissues
with ultrasound,” Journal of medical ultrasonics, vol. 29, no. 4, pp. 155, 2002.
[63] Y. K. Mariappan, K. J. Glaser, and R. L. Ehman, “Magnetic resonance elastography: a review,”
Clinical anatomy, vol. 23, no. 5, pp. 497-511, 2010.
[64] R. Muthupillai, D. Lomas, P. Rossman et al., “Magnetic resonance elastography by direct
visualization of propagating acoustic strain waves,” science, vol. 269, no. 5232, pp. 1854-1857,
1995.
[65] C. Li, G. Guan, Z. Huang et al., “Noncontact all-optical measurement of corneal elasticity,” Optics
letters, vol. 37, no. 10, pp. 1625-1627, 2012.
[66] J. Li, S. Wang, M. Singh et al., "Air-puff OCE for assessment of mouse cornea in vivo." p. 893004.
[67] Y. Qu, Y. He, Y. Zhang et al., “Quantified elasticity mapping of retinal layers using synchronized
acoustic radiation force optical coherence elastography,” Biomedical optics express, vol. 9, no. 9,
pp. 4054-4063, 2018.
[68] Y. Qu, T. Ma, Y. He et al., “Acoustic radiation force optical coherence elastography of corneal
tissue,” IEEE Journal of Selected Topics in Quantum Electronics, vol. 22, no. 3, pp. 288-294, 2016.
[69] Z. Du, R. Li, X. Qian et al., “Quantitative confocal optical coherence elastography for evaluating
biomechanics of optic nerve head using Lamb wave model,” Neurophotonics, vol. 6, no. 4, pp.
041112, 2019.

107

[70] S. Song, N. M. Le, R. K. Wang et al., "Quantitative shear wave optical coherence elastography
(SW-OCE) with acoustic radiation force impulses (ARFI) induced by phase array transducer." p.
93270U.
[71] Ł. Ambroziński, I. Pelivanov, S. Song et al., “Air-coupled acoustic radiation force for non-contact
generation of broadband mechanical waves in soft media,” Applied physics letters, vol. 109, no.
4, pp. 043701, 2016.
[72] Ł. Ambroziński, S. Song, S. J. Yoon et al., “Acoustic micro-tapping for non-contact 4D imaging of
tissue elasticity,” Scientific reports, vol. 6, pp. 38967, 2016.
[73] S. A. M. Tofail, E. P. Koumoulos, A. Bandyopadhyay et al., “Additive manufacturing: scientific and
technological challenges, market uptake and opportunities, ” Materials Today, vol. 21, no. 1,
pp. 22-37, 2018.
[74] N. Shahrubudin, T. C. Lee, and R. Ramlan, “An overview on 3D printing technology:
Technological, materials, and applications,” Procedia Manufacturing, vol. 35, pp. 1286-1296,
2019.
[75] B. Mueller, “Additive manufacturing technologies–Rapid prototyping to direct digital
manufacturing,” Assembly Automation, 2012.
[76] Z.-X. Low, Y. T. Chua, B. M. Ray et al., “Perspective on 3D printing of separation membranes and
comparison to related unconventional fabrication techniques,” Journal of membrane science,
vol. 523, pp. 596-613, 2017.
[77] B. C. Gross, J. L. Erkal, S. Y. Lockwood et al., "Evaluation of 3D printing and its potential impact
on biotechnology and the chemical sciences," ACS Publications, 2014.
[78] J. F. Destino, N. A. Dudukovic, M. A. Johnson et al., “3D printed optical quality silica and silica–
titania glasses from sol–gel feedstocks,” Advanced Materials Technologies, vol. 3, no. 6, pp.
1700323, 2018.
[79] M. Li, Y. Yang, F. Iacopi et al., “3D-printed low-profile single-substrate multi-metal layer
antennas and array with bandwidth enhancement,” IEEE Access, vol. 8, pp. 217370-217379,
2020.
[80] J. E. Huddy, M. S. Rahman, A. B. Hamlin et al., “Transforming 3D-printed mesostructures into
multimodal sensors with nanoscale conductive metal oxides,” Cell Reports Physical Science, vol.
3, no. 3, pp. 100786, 2022.
[81] A. F2792-12a, "Standard terminology for additive manufacturing technologies," ASTM
International, 2012.
[82] D. Hong, D.-T. Chou, O. I. Velikokhatnyi et al., “Binder-jetting 3D printing and alloy development
of new biodegradable Fe-Mn-Ca/Mg alloys,” Acta biomaterialia, vol. 45, pp. 375-386, 2016.

108

[83] P. Kunchala, and K. Kappagantula, “3D printing high density ceramics using binder jetting with
nanoparticle densifiers,” Materials & Design, vol. 155, pp. 443-450, 2018.
[84] F. Yang, M. Zhang, and B. Bhandari, “Recent development in 3D food printing,” Critical reviews
in food science and nutrition, vol. 57, no. 14, pp. 3145-3153, 2017.
[85] W. J. Choi, K. S. Hwang, H. J. Kwon et al., “Rapid development of dual porous poly (lactic acid)
foam using fused deposition modeling (FDM) 3D printing for medical scaffold application,”
Materials Science and Engineering: C, vol. 110, pp. 110693, 2020.
[86] M. P. Chae, W. M. Rozen, P. G. McMenamin et al., “Emerging applications of bedside 3D printing
in plastic surgery,” Frontiers in surgery, vol. 2, pp. 25, 2015.
[87] S. A. Skoog, P. L. Goering, and R. J. Narayan, “Stereolithography in tissue engineering,” Journal of
Materials Science: Materials in Medicine, vol. 25, no. 3, pp. 845-856, 2014.
[88] J. Wang, A. Goyanes, S. Gaisford et al., “Stereolithographic (SLA) 3D printing of oral modified-
release dosage forms,” International journal of pharmaceutics, vol. 503, no. 1-2, pp. 207-212,
2016.
[89] F. P. Melchels, J. Feijen, and D. W. Grijpma, “A review on stereolithography and its applications
in biomedical engineering,” Biomaterials, vol. 31, no. 24, pp. 6121-6130, 2010.
[90] F. Kotz, A. S. Quick, P. Risch et al., “Two ‐Photon Polymerization of Nanocomposites for the
Fabrication of Transparent Fused Silica Glass Microstructures, ” Advanced Materials, vol. 33,
no. 9, pp. 2006341, 2021.
[91] T. Gissibl, S. Thiele, A. Herkommer et al., “Two-photon direct laser writing of ultracompact multi-
lens objectives,” Nature photonics, vol. 10, no. 8, pp. 554-560, 2016.
[92] A. J. D. Shaikeea, H. Cui, M. O’Masta et al., “The toughness of mechanical metamaterials,”
Nature materials, vol. 21, no. 3, pp. 297-304, 2022.
[93] Z. Xu, R. Hensleigh, N. J. Gerard et al., “Vat photopolymerization of fly-like, complex micro-
architectures with dissolvable supports,” Additive Manufacturing, vol. 47, pp. 102321, 2021.
[94] V. Karageorgiou, and D. Kaplan, “Porosity of 3D biomaterial scaffolds and osteogenesis,”
Biomaterials, vol. 26, no. 27, pp. 5474-5491, 2005.
[95] S. J. Hollister, R. Maddox, and J. M. Taboas, “Optimal design and fabrication of scaffolds to
mimic tissue properties and satisfy biological constraints,” Biomaterials, vol. 23, no. 20, pp.
4095-4103, 2002.
[96] R. Wodnicki, H. Kang, R. Chen et al., “Co-integrated PIN-PMN-PT 2-D array and transceiver
electronics by direct assembly using a 3-D printed interposer grid frame,” IEEE transactions on
ultrasonics, ferroelectrics, and frequency control, vol. 67, no. 2, pp. 387-401, 2019.

109

[97] R. Wodnicki, H. Kang, R. Zhang et al., "PIN-PMN-PT single crystal composite and 3D printed
interposer backing for ASIC integration of large aperture 2D array." pp. 1-4.
[98] I. Pelivanov, L. Gao, J. Pitre et al., “Does group velocity always reflect elastic modulus in shear
wave elastography?,” Journal of biomedical optics, vol. 24, no. 7, pp. 076003, 2019.
[99] M. D. Twa, J. Li, S. Vantipalli et al., “Spatial characterization of corneal biomechanical properties
with optical coherence elastography after UV cross-linking,” Biomedical optics express, vol. 5,
no. 5, pp. 1419-1427, 2014.
[100] R. Wodnicki, H. Kang, Y. Sun et al., "High Frequency 1.75 D array using a 3D printed pitch-
changing interposer backing." pp. 1-4.
[101] X. Qian, H. Kang, R. Li et al., “In vivo visualization of eye vasculature using super-resolution
ultrasound microvessel imaging,” IEEE Transactions on Biomedical Engineering, vol. 67, no. 10,
pp. 2870-2880, 2020.

Abstract (if available)

Abstract 2D array fabrication has been an exciting research direction in the past 30 years. Connecting high density element matrix to the system has been one of the main challengings that limit the development of a fully sampled 2D array. The rapid development in 3D printing technology in recent years provides an option to address this challenge.
In this work, we implemented 3D printed interposers with straight channels and progressed to interposers with pitch-shifting structures in fully sampled 2D array fabrication and designed and built different types of 2D arrays targeted at different applications: 2D array with central opening geometry for acoustic radiation force optical coherence elastography (ARF-OCE) and high-frequency (>10MHz) imaging array. ARF-OCE has been successfully implemented to characterize the biomechanical properties of soft tissues such as the cornea and the retina with high resolution using single-element ultrasonic transducers for ARF excitation. In this study, we combined the advantages of 3D dynamic electronic steering of the 2D ultrasonic array and high-resolution optical coherence tomography (OCT) and propose a new method called 2D ultrasonic array-based optical coherence elastography imaging. And 2D imaging array, which enabled real-time volumetric imaging, has shown its great potential in transesophageal echocardiography (TEE) and intracardiac echocardiography (ICE). The performance of the 2D arrays’ elements was measured in pulse-echo(P/E) test and the Ring array’s 3D steering capability was first validated using a hydrophone. The combined 2D ultrasonic array OCE system was calibrated using a homogenous phantom, followed by an experiment on ex vivo rabbit corneal tissue. And the high-frequency 2D array was applied in the imaging test and real-time 3D images were acquired. The results demonstrate that the 3D printed interpose offers another possibility to address the limitation of high-density interconnection and strict geometry requirements in 2D array design and fabrication.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Fabrication of ultrasound transducer and 3D-prinitng ultrasonic device

PDF

High resolution elastography in ophthalmology

PDF

High-frequency ultrasonic transducers for photoacoustic applications

PDF

Design and development of ultrasonic array transducers for specialized applications

PDF

High frequency ultrasound elastography and its biomedical applications

PDF

High frequency ultrasonic phased array system and its applications

PDF

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

PDF

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

PDF

Multi-modality intravascular imaging by combined use of ultrasonic and opticial techniques

PDF

Array transducers for high frequency ultrasound imaging

$Metasurfaces in 3D applications: multiscale stereolithography and inverse design of diffractive optical elements for structured light$

PDF

Metasurfaces in 3D applications: multiscale stereolithography and inverse design of diffractive optical elements for structured light

PDF

Additive manufacturing of piezoelectric and composite for biomedical application

PDF

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

PDF

Accurate 3D model acquisition from imagery data

PDF

Intravascular imaging on high-frequency ultrasound combined with optical modalities

PDF

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

PDF

Engineering 2D & 3D microphysiological systems for interrogating skeletal muscle tissues

PDF

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

PDF

Multi-scale biomimetic structure fabrication based on immersed surface accumulation

Asset Metadata

Creator Kang, Haochen (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Degree Conferral Date 2022-08

Publication Date 08/01/2022

Defense Date 05/19/2022

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

Tag 2D ultrasonic array,3D printing,elastography,OAI-PMH Harvest,ultrasound imaging

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Zhou, Qifa (committee chair), Chen, Yong (committee member), Shen, Keyue (committee member)

Creator Email haochenk@usc.edu,kanghc9355@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111375911

Unique identifier UC111375911

Legacy Identifier etd-KangHaoche-11051

Document Type Dissertation

Format application/pdf (imt)

Rights Kang, Haochen

Type texts

Source 20220801-usctheses-batch-965 (batch), 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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

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

Repository Email cisadmin@lib.usc.edu