Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications
(USC Thesis Other)
Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
AUDIO AND ULTRASOUND MEMS BUILT ON PZT SUBSTRATE FOR
LIQUID AND AIRBORNE APPLICATIONS
by
Youngki Choe
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
(ELECTRICAL ENGINEERING)
December 2012
Copyright 2012 Youngki Choe
ii
To our parents.
iii
Acknowledgements
I would like to express my deepest gratitude to my advisor, Prof. Eun Sok Kim,
for his support and advice that assure the success of this thesis. Without Dr. Kim’s
leadership and insightful instructions, this work wouldn’t be completed.
I am also grateful to Prof. K. Kirk Shung and Prof. John Choma for reading
this thesis and their advice on this thesis. I also appreciate Changyang Lee in Prof.
Shung’s group for spending his valuable time in hydrophone measurement.
I am indebted to all the members in the USCMEMS group for their
camaraderie and support. My deepest gratitude goes to Dr. Shih-jui Chen, who
trained me for lots of equipment when I first joined this group and gave me
tremendous helps. Much gratitude goes to Dr. Anderson Lin and Dr. Hongyu Yu,
who provided lots of advices regarding design and fabrication on acoustic ejectors.
Special thanks to Chuang Yuan Lee for his willingness to provide indispensable
advice and support. I also want to thank my colleagues, Lingtao Wang, Lukas
Baumgartel, Qian Zhang, Arash Vafanejad, and Yufeng Wang for many useful
discussions.
Finally, but most importantly, I would like to give my sincere thanks to my
wife, Hyun Jin Cho, for her support and encouragement and my loving daughter,
Elie Sunhyo Choe. Also, I appreciate my parents for their sacrifice and endless love.
iv
Table of Contents
Acknowledgements .................................................................................................... iii
List of Tables ............................................................................................................. vii
List of Figures .......................................................................................................... viii
Abstract ..................................................................................................................... xiv
Chapter 1
Introduction .................................................................................................................. 1
1.1 Review of PZT Bimorph Microspeakers ............................................. 1
1.2 Review of Valveless Micropumps ....................................................... 4
1.3 Review of Peptide Synthesis using Acoustic Ejector .......................... 5
1.4 Review of Ejector Arrays of 8 Directional Acoustic Ejectors ............. 7
1.5 Review of Acoustic Tweezers .............................................................. 9
1.6 Overview of the Chapters................................................................... 10
Chapter 1 References ..................................................................................... 12
Chapter 2
Hi-Fidelity Microspeaker on PZT Bimorph Diaphragm ............................................ 18
2.1 Transducer Design ............................................................................. 18
2.2 Fabrication ......................................................................................... 21
2.2.1 Bimorph Formation .................................................................... 21
2.2.2 Process Flow for the PZT Bimorph Microspeakers ................... 23
2.3 Experimental Setup ............................................................................ 25
2.4 Measurement Results ......................................................................... 27
2.4.1 Mechanical response of microspeakers ...................................... 27
2.4.2 Acosutic response of microspeakers .......................................... 30
2.4.3 Power consumption measurement of microspeakers ................. 32
2.5 Summary ............................................................................................ 34
Chapter 2 References ..................................................................................... 35
v
Chapter 3
Valveless Micropump Driven by Acoustic Streaming............................................... 36
3.1 Acoustic Fresnel Lens Design ............................................................ 37
3.2 Fabrication ......................................................................................... 39
3.3 Measurement Setup ............................................................................ 41
3.3.1 Micropump with U-shaped channel ........................................... 41
3.3.2 Micropump with straight-lined channel ..................................... 42
3.4 Measurement Results ......................................................................... 44
3.4.1 Micropump with U-shape channel ............................................. 44
3.4.2 Micropump with straight-lined channel without reflector ......... 45
3.4.3 Micropump with straight-lined channel with reflector .............. 50
3.5 Summary ............................................................................................ 53
Chapter 3 References ..................................................................................... 54
Chapter 4
Peptide Synthesis on a Modified Glass Surface Using Acoustic Droplet Ejector ..... 55
4.1 Materials for Peptide Synthesis .......................................................... 56
4.2 SPOT Protocol to Synthesize Peptides on a Glass Substrate ............. 56
4.3 Ejector Design and Fabrication .......................................................... 59
4.4 Measurement Setup ............................................................................ 61
4.5 Measurement Results ......................................................................... 63
4.5.1 Optimum coupling time measurement ....................................... 63
4.5.2 Ejecting condition calibration .................................................... 65
4.5.3 Peptide synthesis using ejectors ................................................. 66
4.6 Summary ............................................................................................ 71
Chpater 4 References ..................................................................................... 72
Chapter 5
On-Chip Integration of Eight Directional Droplet Ejectors to Ink a Spot ................. 73
5.1 Working Principle .............................................................................. 73
5.2 Transducer Design ............................................................................. 76
5.3 Fabrication and Test Setup ................................................................. 80
5.4 Measurement Results and Discussions .............................................. 83
5.5 Summary ............................................................................................ 86
Chpater 5 References ..................................................................................... 87
vi
Chapter 6
Acoustic Tweezers with Multi-foci Fresnel Lens ...................................................... 88
6.1 Negative Axial Radiation Force in a Bessel Beam ............................ 88
6.2 Lens Designs ...................................................................................... 90
6.2.1 Aluminum axicon lens ............................................................... 90
6.2.2 Multi-foci Fresnel lens ............................................................... 90
6.3 Device Fabrication ............................................................................. 92
6.3.1 Acoustic tweezers with aluminum axicon lens .......................... 92
6.3.2 Acosutic tweezers with multi-foci Fresnel lens ......................... 94
6.4 Particle trapping measurement setup and results ............................... 95
6.4.1 Measurement setup .................................................................... 95
6.4.2 Lipid particle trapping ................................................................ 97
6.4.3 Polystyrene microparticle trapping ............................................ 99
6.4.4 Beam profile measurement ...................................................... 101
6.4.5 Zebra fish egg trapping ............................................................ 104
6.5 Summary .......................................................................................... 106
Chpater 6 References ................................................................................... 107
Chapter 7
Conclusion and Future Directions ............................................................................ 108
7.1 Thinner Microspeaekrs on Round PZT Diaphragm ......................... 108
7.2 Micropumps with Non-vertical Reflector ........................................ 109
7.3 Protein Synthesis with Ejector Array ............................................... 110
7.4 Higher Harmonic Acoustic Tweezers .............................................. 110
Bibliography ............................................................................................................. 112
vii
List of Tables
Table 3.1 Particle drift velocity of the micropump with straight-lined
channel .................................................................................................. 46
Table 5.1 Ejection condition for the eight directional ejectors of the
ejector array ........................................................................................... 84
viii
List of Figures
Figure 2.1 Contour plots of the normal stresses (a)
xx
and (b)
yy
....................... 20
Figure 2.2 Contour plots of (a) the area where the normal stresses
xx
and
yy
have opposite sign and of (b) the summed normalized
stress (
xx
+
yy
) for those areas where
xx
and
yy
have same
sign ........................................................................................................ 20
Figure 2.3 The cross-sectional view of stress distributions on a four-edge-
clamped bimorph diaphragm when a uniform loading is
applied to the diaphragm. ...................................................................... 21
Figure 2.4 SEM photos of the cross-section of the fabricated PZT
bimorph diaphragm with (a) and without (b) the mechanical
polishing. ............................................................................................... 23
Figure 2.5 Brief fabrication steps for a PZT-based bimorph microspeaker ........... 24
Figure 2.6 Photos of the fabricated microspeaker taken from (a) the
backside and from (b) the front side. .................................................... 25
Figure 2.7 Testing setup for the displacement measurement of the
acoustic transducer ................................................................................ 26
Figure 2.8 Testing setup for the sound pressure output measurement of
the acoustic transducer .......................................................................... 26
Figure 2.9 Measured displacements of the microspeaker with patterned
top electrode and un-patterned bottom electrode .................................. 29
Figure 2.10 Measured displacements of the microspeaker with patterned
top and bottom electrode ....................................................................... 29
Figure 2.11 Sound pressure outputs measured at 5 mm away from the
microspeaker having patterned top electrode and un-patterned
bottom electrode with a 0.7 cc cylindrical encapsulating
package .................................................................................................. 31
Figure 2.11 Sound pressure outputs measured at 5 mm away from the
microspeaker having patterned top and bottom electrodes with
a 0.7 cc cylindrical encapsulating package ........................................... 31
ix
Figure 2.13 Power measurement setup ..................................................................... 32
Figure 2.14 Measured real power consumption of the PZT bimorph
microspeaker and a commercial cell-phone speaker ............................. 33
Figure 3.1 MATLAB simulation results showing in-plane particle
displacements on the focal plane of the Fresnel lens composed
of (a) 5 outer rings and (b) all possible rings. The yellow pie-
shaped annular lines represent Fresnel lens shape ................................ 37
Figure 3.2 Arrangement of the sectored Fresnel lenses in U-shaped fluidic
channel. The body force due to the focused acoustic beam is
designed to be directed along the line tangential to the curved
shape of the U-shaped channel .............................................................. 38
Figure 3.3 Brief fabrication steps of the micropumps ............................................ 39
Figure 3.4 Photos of the fabricated micropumps with U-shaped channels:
(a) the micropump built on the sectored Fresnel lens shown in
Figure 3-1b and (b) the micropump built on the Fresnel lens
utilizing only 5 outer rings of the Fresnel lens shown in Figure
3-1a ........................................................................................................ 40
Figure 3.5 Photos of the fabricated micropumps having straight-line
channel (a) without and (b) with an acrylic reflector ............................ 40
Figure 3.6 Measurement setup for the micropump with U-shaped channel ........... 41
Figure 3.7 Photo from top view of the setup to measure pumping rate of
the micropump: (a) a reservoir for DI water, (b) an inline
micropump, (c) a water drain tube placed on the same level
with the liquid channel to minimize the differential pressure
between the reservoir and the drain tube ............................................... 43
Figure 3.8 Photo from side view of the setup to measure pumping rate of
the micropump ...................................................................................... 43
Figure 3.9 Photos (from a video) showing that a group of polystyrene
microspheres were moved about 7.3 mm in 100 msec ......................... 44
Figure 3.10 Photos showing width of fast water stream according to
different channel width: (a) 6 mm, (b) 8 mm, and (c) 10 mm .............. 46
x
Figure 3.11 Measured particle drift velocities vs. pulse repetition frequency
(PRF) as a function of the water height ................................................ 47
Figure 3.12 MATLAB simulation results showing in-place particle
displacement observed at 2 mm away from acoustic Fresnel
lens which has focal length of 4 mm ..................................................... 49
Figure 3.13 MATLAB simulation results showing in-place particle
displacement observed at 6 mm away from acoustic Fresnel
lens which has focal length of 4 mm ..................................................... 49
Figure 3.14 A conceptual diagram showing the working principle of the
acrylic reflector ..................................................................................... 49
Figure 3.15 Plot of pulse repetition frequency (PRF) vs. particle drift
velocity of the micropump with acrylic reflector measured with
different distances between the micropump and the liquid
surface ................................................................................................... 52
Figure 3.16 Plot of applied power vs. particle drift velocity of the
micropump with acrylic reflector with different distances
between the micropump and the liquid surface ..................................... 52
Figure 4.1 Preparation of amine-terminated glass surface ..................................... 57
Figure 4.2 Conceptual diagrams showing amino acid coupling and FITC
attachment procedures ........................................................................... 58
Figure 4.3 Fabrication steps of an acoustic ejector with a silicon lens .................. 60
Figure 4.4 Photos of fabricated device: (a) top view and (b) bottom view
of the packaged ejector. (c) an ejector without top cover. .................... 60
Figure 4.5 Measurement setup of pre-activated amino acid solution
ejection onto a slide glass ...................................................................... 62
Figure 4.6 Photo of the fluorescent microscope ..................................................... 62
Figure 4.7 Optical detection of FICT using a fluorescent microscope ................... 63
Figure 4.8 Normalized light intensities measured from active spots with
various coupling times .......................................................................... 64
Figure 4.9 Photo of the pre-activated amino acid solution droplet ejection ........... 65
xi
Figure 4.10 Collected droplets of pre-activated amino acid solution ejected
by the acoustic droplet ejector with silicon lens. Each liquid
bump has 60 nL of the solution ............................................................. 66
Figure 4.11 Ladder structure of peptide chain .......................................................... 68
Figure 4.12 FITC attached at the end of the peptide chain after synthesizing
desired number of peptides ................................................................... 68
Figure 4.13 Photos showing droplet’s position at (a) 150 μsec, (b) 450 μsec,
(c) 700 μsec, and (d) 1000 μsec after the droplet is ejected
from the ejection chamber ..................................................................... 69
Figure 4.14 Fluorescent images taken from the 9 spots in the peptide ladder
structure ................................................................................................. 69
Figure 4.15 Plot of measured light intensity............................................................. 70
Figure 5.1 MATLAB simulation results showing (a) vertical and (b) in-
plane direction particle displacement due to focused acoustic
beam generated from a transducer with circular top and bottom
electrode ................................................................................................ 74
Figure 5.2 MATLAB simulation results showing (a) vertical and (b) in-
plane direction particle displacement due to the focused
acoustic beam generated from a transducer with pie-shape top
and bottom electrode with 90 ° apex angle ........................................... 75
Figure 5.3 Photo of an obliquely ejected droplet after 400 μsec from the
ejection .................................................................................................. 75
Figure 5.4 Schematic diagram of the array of eight directional ejectors for
inking a spot without moving the ejectors ............................................ 77
Figure 5.5 Photo of 8 Fresnel lenses sitting on circumference of a circle
on the fabricated PZT transducer .......................................................... 77
Figure 5.6 Photos of the fabricated PZT ejector array with silicon
microchannels etched by deep reactive ion etching (DRIE). ................ 78
Figure 5.7 Photos of the Fresnel lenses of the ejectors after the silicon
wafer with DRIE-etched microchannels was glued to the lens-
containing PZT substrate ....................................................................... 79
xii
Figure 5.8 Brief fabrication steps of the ejector array ............................................ 81
Figure 5.9 Measurement setup of ejector array’s inking a spot operation ............. 82
Figure 5.10 Photos of (a) top-view and (b) bottom-view of the fabricated
ejector .................................................................................................... 82
Figure 5.11 Photos of droplet ejection process captured at (a) 50 μsec, (b)
300 μsec, (c) 900 μsec, and (d) 2,000 μsec after actuating
Ejector 1 and Ejector 2 .......................................................................... 83
Figure 5.12 Photo of the water bump collected on a glass slide after the 8
ejectors of the ejector array ejected about 200 droplets onto the
glass slide. Red dots indicate the center of the water bumps ................ 85
Figure 5.13 Plot of the center locations of the water bumps (on the glass
slide) formed by the droplets ejected by the 8 ejectors without
moving the ejectors, showing close proximity of the droplet
placements ............................................................................................. 85
Figure 6.1 Schematics of the aluminum alloy axicon lens: (a) top view (b)
side view................................................................................................ 90
Figure 6.2 Ray trajectories of the acoustic waves going through an axicon
lens ........................................................................................................ 91
Figure 6.3 Side-view schematic of the multi-foci Fresnel lens. ............................. 92
Figure 6.4 Fabrication steps of the acoustic tweezers with an aluminum
alloy axicon lens. ................................................................................... 93
Figure 6.5 Photos of (a) aluminum alloy axicon lens (b) acoustic tweezers
with the fabricated aluminum alloy axicon lens ................................... 93
Figure 6.6 (a) SEM photo of the fabricated multi-foci Fresnel lens on a
PZT sheet (b) Photo of a packaged acoustic tweezers with the
multi-foci Fresnel lens........................................................................... 94
Figure 6.7 Brief fabrication steps of the acoustic tweezers built on a 127
μm thick PZT sheet with multi-foci Fresnel lens. ................................. 95
Figure 6.8 Schematic of the setup to observe the trapping of
microparticles by the fabricated acoustic tweezers ............................... 96
xiii
Figure 6.9 Photo of manual XYZ control. .............................................................. 97
Figure 6.10 Photos taken at different times showing that a trapped lipid
particle holds its position even when hit by another moving
lipid particle .......................................................................................... 98
Figure 6.11 A large lipid particle with 200 μm diameter was trapped by the
fabricated acoustic tweezers. ................................................................. 98
Figure 6.12 Photo of trapped microparticle with 80 μm diameter by an
acoustic tweezers with a silicon window ............................................ 100
Figure 6.13 A group of polystyrene microspheres with 80 μm diameter was
trapped by an acoustic tweezers without a silicon window ................ 100
Figure 6.14 Photo of hydrophone measurement setup (a) projection view (b)
side view.............................................................................................. 101
Figure 6.15 Hydrophone measurement results along z-axis (wave
propagation direction) ......................................................................... 102
Figure 6.16 Hydrophone measurement results along lateral direction at 2.9
mm away from the acoustic tweezers ................................................. 103
Figure 6.17 Hydrophone measurement results along lateral direction at 4.0
mm away from the acoustic tweezers ................................................. 103
Figure 6.18 Hydrophone measurement results along lateral direction at 5.8
mm away from the acoustic tweezers. ................................................ 103
Figure 6.19 Two polystyrene microparticles with diameter of 500 μm are
trapped by the acoustic tweezers. ........................................................ 105
Figure 6.20 A zebra fish egg was successfully trapped by the acoustic
tweezers with multi-foci Fresnel lens. ................................................ 105
xiv
Abstract
This thesis presents piezoelectric microelectromechanical systems (MEMS) for
sound generation, liquid pumping, peptide synthesizing, and microparticle trapping.
The common thread for the piezoelectric MEMS is the usage of piezoelectric lead
zirconate titanate (PZT) sheet in generating sound waves for various applications in
air and liquid.
For air-borne audio sound generation, a microspeaker based on bending
movement of PZT bimorph diaphragm was designed and fabricated. A PZT substrate
was mechanically polished and glued to a micromachined silicon with low viscosity
glue to minimize the glue thickness for minimum electric field drop in the glue. Top
and bottom electrodes of the microspeaker were patterned according to the stress
distribution of a diaphragm for uniform loading to optimize the piezoelectric
actuation of bimorph structure. With an order of magnitude lower power
consumption compared to a commercial cell-phone electromagnetic microspeaker,
the newly fabricated microspeaker produced about 20 dB higher sound output in
frequency range of 500 ~ 2,000 Hz. The power consumption increased as the
frequency increased, but the sound output was still much higher than a commercial
cell phone microspeaker.
For liquid pumping, 260 μm-thick PZT sheet was actuated on its fundamental
resonance frequency to generate 8.6 MHz acoustic waves. Acoustic Fresnel lenses
were arranged along with U-shape and straight-line liquid channel to produce
xv
unidirectional liquid flow. The sector angle of pie-shape electrodes and Fresnel lens
were optimized to generate larger in-plane direction acoustic body force, which
directly affect the liquid flow rate in the channel. To improve the pumping rate and
stabilize the operation condition, acrylic reflector was set up between Fresnel lenses.
The micropump with U-shaped lens arrangement produced 7.3 cm/sec particle drift
velocity on the liquid surface, and the micropump having straight-lined lens arranged
and acrylic reflector showed pumping rate of 9.5 mL/min.
For peptide synthesis using the Spot technique, acoustic droplet ejector with
silicon lens structure was used to dispense pre-activated amino-acid on modified
glass surface. Conventional acoustic ejector employing Fresnel lens with air-reflect
uses parylene D polymer film as its structural material. The parylene acoustic Fresnel
lens structure layer was replaced with bare silicon not only to simplify the fabrication
steps but also to improve the durability of the device and convenience of washing out
solution inside the ejector. With surface modified glass substrate and acoustic ejector
as a liquid dispenser, 9-mers long peptide was synthesized with 70% of final
synthesis yield that is equivalent with 96% stepwise synthesis yield.
To expand the functionality of the acoustic droplet ejector as a liquid dispenser
in peptide synthesizing system, an ejector array consisting of eight directional
ejectors were designed and fabricated. The Fresnel lenses and electrodes of eight
transducers were patterned into pie-shape with apex angle of 90° to reduce the
distances between ejectors in an ejector array. The silicon wafer was etched with
xvi
Deep Reactive Ion Etching (DRIE), and combined to the PZT transducer, with a
mask aligner aiding on the alignment, to form ejecting chambers and liquid transfer
channels. The chambers were designed to provide identical acoustic echo
circumstances for each ejector in the array. The six ejectors out of the eight ejectors
in an array were able to eject droplets onto a very small area of 238 x 380 μm
2
when
actuated individually.
For particle trapping, an acoustic tweezers was designed and fabricated on 128
μm-thick PZT sheet. A Fresnel lens with multiple focal lengths was designed to
generate an acoustic Bessel beam, similar to what axicon lens does. A multi-foci
Fresnel was designed to have focal lengths of 830 μm, 860 μm, and 890 μm for its
two most inner rings, next two rings, and the remaining 3 rings, respectively, out of
total 7 Fresnel lens rings. The acoustic tweezers with the designed multi-foci Fresnel
lens showed very good trapping ability on lipids particles, polystyrene microparticles,
and zebra fish eggs, and confirmed the generation of an acoustic Bessel beam.
1
Chapter 1
Introduction
Piezoelectric lead zirconate titanate (PZT) sheet when combined with
micromachined silicon wafer opens up many opportunities to innovate in sound
wave generation and its applications. This thesis presents bimorph microspeakers
(chapter 2), valve-less micropumps (chapter 3), peptide synthesis using acoustic
droplet ejector (chapter 4), ejector arrays of 8 directional acoustic ejectors (chapter 5),
and acoustic tweezers (chapter 6), all of which are based on piezoelectric
microelectromechanical systems (MEMS) built on PZT sheet. General reviews of the
five areas are covered in this chapter.
1.1 Review of PZT Bimorph Microspeakers
With rapid market expansion of portable electronics, demands for
microspeakers having high fidelity, large sound output, small volume, and low
power consumption are emerging dramatically. A lot of researches were performed
to miniaturize conventional electromagnetic loud speakers consisting of permanent
magnets and voice coils and to implement them in cell phones [1-5]. In spite of
relatively larger size of electromagnetically actuated microspeakers, the sound output
was not impressive even after spending lots of efforts on optimizing voice coil [6]
and diaphragm [7].
2
A Microelectromechanical Systems (MEMS) technology was introduced to
make smaller microspeakers and to improve the energy convergence efficiency from
electrical energy to acoustic energy. Various actuation methods such as
electromagnetic actuation same as conventional loud speakers, electrostatic actuation
[8-9], and piezoelectric actuation was tried. Microspeakers using electromagnetic
actuation were still bulky compared to microspeakers using other actuation methods
because it needs to have permanent magnetic which is hard to miniaturize. In
addition to the larger size, the microspeakers with electromagnetic actuation
consume higher power [10-11] than the other two types of microspeakers because
relatively large amount of current flows in the voice coil. However, the
microspeakers using electromagnetic actuation requires much lower voltage to
actuate the transducer.
Piezoelectric transduction requires high voltage to obtain large displacement.
However, the power consumption is relatively small compared to electromagnetic
transduction because very little amount of current flows through transducers due to
high dielectric constant of piezoelectric material. Since three edges of a rectangular
diaphragm released in a cantilever structure, piezoelectric cantilever microspeakers
can have large displacement at the end the cantilever [12-13]. However, the air gap
between a transducer and the released cantilever degrades the sound generation
efficiency because of an acoustic short circuit effect. To maximize the electric-to-
acoustic transduction efficiency, it is very important to have flat cantilever and to
3
have minimum air gap when the transducer is not actuated. The thickness direction
residual stress was optimized in microspeakers with ZnO layer as a piezoelectric
material to have flat cantilever [12], and a PZT was tried in another
microphone/microspeaker to have better sensitivity and sound output utilizing PZT’s
high piezoelectric constant [13].
Diaphragm based piezoelectric microspeakers were also fabricated and tested
[14-15]. Since four edges are clamped in diaphragm based acoustic transducers,
diaphragm based microspeakers are much less influenced by the acoustic short
circuit effect. Therefore, microspeakers on a diaphragm can generate large sound
pressure once the device generates significant displacements. However, since there is
damping due to the mechanical coupling between diaphragm and transducer body
frame, large actuation force is needed to induce displacement.
A bimorph diaphragm can effectively generate bending displacements in four
edge clamped diaphragm. To actuate piezoelectric bimorph diaphragm, both top and
bottom electrodes need to be patterned with very good alignment to each other. In
this work, we present microspeakers built on a PZT bimorph diaphragm that is
supported by bulk-micromachined silicon body. The top and bottom electrodes
sandwiching 127 μm thick PZT were designed to maximize the bending momentum
in a given electric field. Two squared PZT sheets were aligned along their two
sidelines with good enough alignment accuracy to generate large sound output. The
fabricated transducers generated at least 20 dB larger sound outputs than
4
conventional microspeakers having very thin thickness (the device thickness is less
than 700 μm). The power consumption was turned out to be an order of magnitude
smaller than that of commercial cell phone microspeaker from 100 Hz to 2 kHz.
1.2 Review of Valveless Micropumps
Demands for miniaturized liquid circulation system have significantly
increased for biomedical applications [16-17] and electronics cooling [18-21]. A
micropump either uses a set of valves or is valveless, though the driving mechanisms
include a variety of options including electrostatic, piezoelectric [22-24],
electromagnetic [25-26], thermal (particularly using shape memory alloy [27]), and
etc. Valve-based micropumps offer a large backpressure and a low mixing of up and
down streams, but suffer from a relatively low flow rate and poor long-term
reliability. Also, the fabrication process of valve-based micropumps is inherently
more complicated than that of valveless micropumps. In general, valveless
micropumps can pump liquid at a high flow rate, with a good consistency and long-
term reliability, as long as there is not much backpressure. Thus, valveless
micropumps are very useful in applications where the back-pressure is low and does
not vary much. They are more suitable for circulating liquid around a channel.
This work describes two types of valveless micropumps based on Fresnel
acoustic lens built on a 260 μm-thick PZT substrate: U-shaped micropumps and
straight-lined micropumps. The PZT transducer produces about 8.6 MHz acoustic
5
waves, which are focused onto multiple spots along a liquid channel through a set of
Fresnel lenses. With about 5 mm focal length of the acoustic Fresnel lens, the U-
shaped micropumps produced U-shape liquid flow with 7.3 cm/sec flow speed, and
the straight-lined micropump was capable of producing 159 μL/sec (or 9.5 mL/min)
pumping rate that is more than an order of magnitude improvement compared to
reported pumping rate of other micropumps. The major features of the acoustic PZT
micropumps include (1) actuation by focused acoustic beam, (2) large liquid pump
rate, (c) high long-term reliability, and (4) valveless mechanism.
1.3 Review of Peptide Synthesis using Acoustic Ejector
A classical method to synthesize peptides was liquid-phase peptide synthesis.
The liquid-phase peptide synthesis method is still in use for industrial mass
production. However, solid-phase peptide synthesis (SPPS) that was pioneered by
Robert Bruce Merrifield [28] replaced the liquid-phase peptide synthesis in most
modern labs due to economic reason.
In SPPS, small solid beads (or resins) usually made of porous glass are treated
with linkers on which peptide chain is built. The synthesized peptide stays on the
beads until it is chemically cleaved by a reagent. Since the overall peptide synthesis
yield is exponential function of the stepwise yield of adding single amino acid, it is
very important to maximize the stepwise yield. For example, the final yield
synthesizing 26-mers-long peptide is 77% with 99% stepwise yield, but the final
6
yield would drop to 25% when the stepwise yield decreased to 95%. For this reason,
SPPS chemistry has been significantly optimized to have good stepwise yield for last
50 years in resins [29] and linkers [30-32].
Though SPPS requires smaller volume of chemicals compared to liquid-phase
synthesis, SPPS still consumes much larger volume of chemicals than actual needs in
labs. In addition, SPPS synthesis technique can synthesize one kind of peptide at a
time. The Spot synthesis technique was introduced by Ronald Frank [33] to realize
in-situ parallel peptide synthesis on membrane support. The Spot technique follows
the principle and chemistry of SPPS removing the beads involvement. A Spot
synthesis, in general, forms array of spots on cellulose membrane, and different
sequences of peptides are synthesized on each spot in parallel [34-38].
In the Spot peptide synthesis on cellulose membrane supports, the spot size
varies from several millimeters to few centimeters in diameter because the dispensed
chemicals spread out and, consequently, occupy large area in the membrane. A Spot
synthesis was demonstrated on modified glass surface [39]. Higher density of
peptide array can be formed on slide glass because glass does not observe any liquid.
Since the dispensing volume of chemical should be reduced accordingly as the spot
size decrease, liquid dispensing mechanism with high resolution on volume control
is necessary.
In this work, peptides of glycine with various lengths from 1-mer to 9-mers
were synthesized on modified glass surface using the Spot peptide synthesis protocol.
7
An acoustic ejector employing silicon acoustic Fresnel lens dispensed droplets of
pre-activated amino acid solution onto the glass substrate. The stepwise coupling
yield was calculated to be 96% by measuring the light intensity coming from FITC
fluorescent tag, and the final yield of 9-mers-long peptide synthesis was calculated to
be 70% accordingly.
When precise liquid dispenser with 200 pL volume control resolution is
combined with the Spot protocol, the protein chip density was dramatically increased.
In 25 mm x 75 mm glass slide, 36 active spots having free amine group were defined
to support in-situ parallel peptide synthesis. The protein chip density can be easily
increased further because very generous spacing rule was applied in this work as a
preliminary demonstration. Utilizing this work, in-situ protein chip fabrication with
arbitrary on-demand sequence can be done in much shorter time with good economy.
1.4 Review of Ejector Arrays of 8 Directional Acoustic Ejectors
The advantage of nozzleless acoustic droplet ejectors [40] over nozzle-based
ejectors is versatility due to clog-free ejection and directional ejections [41]. Using
the directional ejection capability, we can arrange multiple acoustic ejectors to ink a
spot with multiple chemicals without replacing or having to move the ejectors. A
single ejector [42] and an ejector array of four ejectors [43] successfully synthesized
DNA sequences on a glass slide with droplets containing DNA bases. Unlike DNA
synthesis for which only 4 kinds of nucleotides (Adenine, Thymine, Cytosine, and
8
Guanine) are needed, protein synthesis on a chip requires much larger variety of
droplets, because there are about 20 kinds of peptides [39]. Thus, a dense ejector
array of directional ejectors capable of inking one spot with 20 different peptides
without having to move the ejectors is highly desired. Recalling the review regards
to Spot array synthesis of peptides, an ejector array can be utilized in in-situ parallel
peptide synthesis on modified glass surface by dispensing different types of pre-
activated amino acids without replacing or moving ejector.
This work describes an array of eight directional acoustic ejectors integrated
with silicon microfluidic channels, chambers, and reservoirs fabricated by Deep
Reactive Ion Etching (DRIE) that enables precise dimension control and high aspect
ratio on silicon. The silicon wafer containing microchannels and reservoirs was
bonded to a lead zirconate titanate (PZT) substrate with a low viscosity glue, and
multiple liquids in the reservoirs were automatically delivered to the liquid chambers
over the PZT ejectors through the microchannels by capillary force (as well as
hydrostatic pressure in some cases).
The fabricated ejector array was able to ink very small area (238 x 380 μm
2
)
with droplets that were ejected from 6 different ejectors in the array. If we include
the droplets ejected from two ejectors which have defect acoustic lens, the inked area
size is slightly increased to 399 x 1080 μm
2
. However, the increased area size is still
acceptable considering the 1 mm diameter of the active spot in tested protein array.
9
1.5 Review of Acoustic Tweezers
Contactless particle trapping has many potential applications in biology,
physical chemistry, and bio-medical research. Among the various approaches,
optical tweezers is the most well developed and commonly investigated technique
[44-46]. However, particle trapping based on optical manipulation has several
limitations: the high energy of a highly focused laser can damage living cells [47,48].
Secondly, the size of the largest particle that can be trapped by an optical tweezers is
limited to tens of micrometers. Finally, the use of optical tweezers is confined to
optically transparent objects or shallow region of opaque medium.
Research into ultrasonic tweezers seeks to provide an alternative to optical
tweezers. Ultrasound is more attractive than light, because it imparts much higher
energy (or impact force) and also offers greater spatial range, consequently providing
the ability to capture larger particle over wider range. However, ultrasonic tweezers
has not been widely pursued, possibly due to lack of good focusing technology. An
acoustical tweezing system was pioneered by Wu [49] with a device capable of
stably capturing latex particles or diameter 270 μm. The tweezers utilized two
counteracting acoustic transducers to create a force potential well that could capture
microparticles. More recently, a highly focused transducer was used to capture a
lipid particle that was already confined on a Mylar sheet in water [50].
Optical tweezers mainly used an optical Bessel beam, which can be generated
by an axicon lens, to generate trapping force in a three dimensional space. An
10
ultrasound Bessel beam is also able to induce strong trapping force in a liquid. In
spite of the great advantages of a Bessel beam such as non-diffractive, self-healing,
and having negative axial pressure, an ultrasound Bessel beam was hardly used in an
acoustic tweezers due to the difficulty of fabricating miniaturized acoustic axicon
lens, because an acoustic axicon lens requires large lens angle, precise angle control,
and very smooth lens surface.
This work describes microparticle trapping on a liquid surface by a single
ultrasonic transducer. We designed and fabricated transducers that could generate a
Bessel beam containing a region of a negative axial radiation force [51]. Though
Bessel beams can be generated by an Axicon lens [52], the fabrication of an Axicon
lens for microparticle trapping can be challenging due to the steep lens angle
required. Thus, we fabricated both an aluminum alloy Axicon lens and a multi-foci
Fresnel lens employing an air-reflector [53] that has wide tolerance to the lens
thickness unlike the stepped plate structure [54] and also has wide tolerance to
manufacturing imprecision. The acoustic tweezers with multi-foci Fresnel lens
showed very strong trapping microparticles, and the trapping was considered as a
major evidence of a Bessel beam, as a positive pressure field expels particles out of
the spot.
1.6 Overview of the Chapters
In Chapter 1, reviews of the current methods used for microspeaker, valveless
micropumps, and peptide synthesis using acoustic ejector, ejector arrays, and
11
acoustic tweezers along with the motivation of the thesis work are described as a
brief introduction to the thesis.
Chapter 2 presents the design, and fabrication of the proposed PZT bimorph
microspeaker. The experimental testing results for both microspeaker with patterned
top and bottom electrodes and the microspeaker with patterned top electrodes only
are also described.
Chapter 3 describes the design, fabrication of micropumps built on 260 μm
thick PZT. The optimization of lens pattern, channel shape, and operating condition
is presented.
Chapter 4 describes the peptide synthesis on modified glass surface using
single acoustic ejector. Fmoc chemistry was used for the Spot peptide synthesis on a
modified glass surface. A ladder structure (1-mer to 9-mer) of glycine peptide array
was demonstrated.
Chapter 5 describes design, simulation, and fabrication of an ejector array
consisting of 8 directional ejectors. Each ejector in the array was actuated
individually or simultaneously to demonstrate ejection accuracy.
Chapter 6 describes design and fabrication of acoustic tweezers with aluminum
alloy axicon lens and multi-foci acoustic Fresnel lens. Polystyrene microparticles,
lipids, and a zebra fish egg were trapped by the acoustic tweezers with multi-foci
Fresnel lens as the evidence of an ultrasound Bessel beam formation.
Finally, Chapter 7 presents conclusions and future research directions.
12
Chapter 1 References
[1] G.Y. Hwang, H.G. Kim, S.M. Hwang, and B.S. Kang, “Analysis of Harmonic
Distortion Due to Uneven Magnetic Field in a Microspeaker Used for Mobile
Phones,” IEEE Transactions on Magnetics, vol. 38, pp. 2376-2378, 2002.
[2] S.M. Hwang, G.Y. Hwang, J.H. Kwon, H.J. Lee, and B.S. Kang, “Performance
Comparison Between Circular and Elliptical Type Microspeakers for Cellular
Phones,” IEEE Transactions on Magnetics, vol. 39, pp. 3256-3258, 2003.
[3] S.M. Hwang, J.H. Kwon, and K.S. Hong, “Development of Woofer
Microspeaker Used for Cellular Phones,” IEEE Transactions on Magnetics, vol.
41, pp. 3808-3810, 2005.
[4] S.M. Hwang, H.J. Lee, K.S. Hong, B.S. Kang, and G.Y. Hwang, “New
Development of Combined Permanent-Magnet Type Microspeakers Used for
Cellular Phones,” IEEE Transactions on Magnetics, vol. 41, pp. 2000-2003,
2005.
[5] P.C.P. Chao, C.W. Chiu, and Y. Hsu-Pang, “Magneto-Electrodynamical
Modeling and Design of a Microspeaker Used for Mobile Phones With
Considerations of Diaphragm Corrugation and Air Closures,” IEEE
Transactions on Magnetics, vol. 43, pp. 2585-2587, 2007.
[6] Y.C. Chen, W.T. Liu, T.Y. Chao, and Y.T. Cheng, “AN OPTIMIZED CU-NI
NANOCOMPOSITE COIL FOR LOW-POWER ELECTROMAGNETIC
MICROSPEAKER FABRICATION,” Transducers ’09, IEEE International
Conference on Solid-State Sensors and Actuators, Denver, CO, June 21-25, pp.
25-28, 2009.
[7] C.M. Lee, J.H. Kwon, K.S. Kim, J.H. Park, and S.M. Hwang, “Design and
Analysis of Microspeakers to Improve Sound Characteristics in a Low
Frequency Range,” IEEE Transactions on Magnetics, vol. 46, pp. 2048-2051,
2010.
[8] H. Kim, A.A. Astle, K. Najafi, L.P. Bernal, P.D. Washabaugh, and F. Cheng,
“Bi-directional Electrostatic Microspeaker with Two Large-Deflection Flexible
Membranes Actuated by Single/Dual Electrodes,” Sensors 2005, IEEE
Conference on Sensors, Irvine, CA, October 31-November 3, pp. 89-92, 2005.
[9] R.C. Roberts, J. Du, A.O. Ong, D. Li, C.Z. Zorman, and N.C. Tien,
“Electrostatically Driven Touch-Mode Poly-SiC Microspeaker,” Sensors 2007,
IEEE Conference on Sensors, Atlanta, GA, October 28-31, pp. 284-287, 2007
13
[10] F.L. Ayatollahi and B.Y. Majlis, “Materials Designed and Analysis of Low-
Power MEMS Microspeaker Using Magnetic Actuation Technology,”
Advanced Materials Research, vol. 74, pp. 243-246, 2009.
[11] Y.C. Chen and Y.T. Cheng, “A LOW-POWER MILLIWATT
ELECTROMAGNETIC MICROSPEAKER USING A PDMS MEMBRANE
FOR HEARING AIDS APPLICATION,” IEEE International Micro Electro
Mechanical Systems Conference, Cancun, Mexico, January 23-27, pp. 1213-
1216, 2011.
[12] S.S. Lee, R.P. Ried, and R.M. White, “Piezoelectric Cantilever Microphone and
Microspeaker,” Journal of Microelectromechanical Systems, vol. 5, pp. 238-242,
1996.
[13] R. Tian-Ling, Z. Lin-Tao, L. Li-Tian, and L. Zhi-Jian, “Design Optimization of
Beam-Like Ferroelectrics-Silicon Microphone and Microspeaker,” IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 49,
pp. 266-270, 2002.
[14] S.C. Ko, Y.C. Kim, S.S. Lee, S.H. Choi, and S.R. Kim, “Micromachined
piezoelectric membrane acoustic device,” Sensors and Actuators A: Physical,
vol. 103, pp. 130-134, 2003.
[15] H.J. Kim, K. Koo, S.Q. Lee, K.H. Park, and J. Kim, “High Performance
Piezoelectric Microspeakers and Thin Speakers Array System,” ETRI Journal,
vol. 31, pp. 680-687, 2009.
[16] S. Guo and J. Wang, “A Novel Type of Micropump Using Solenoid Actuator
for Biomedical Applications,” 2007 IEEE International Conference on Robotics
and Automation, Roma, Italy, April 10-14, pp. 654-659, 2007.
[17] G.H. Kwon, G.S. Jeong, J.Y. Park, J.H. Moon, and S.H. Lee, “A low-energy-
consumption electroactive valveless micropump for long-term biomedical
applications,” Lab on a Chip, vol. 11, pp. 2910-2915, 2011.
[18] V. Singhal, “A novel valveless micropump with electrohydrodynamic
enhancement for high heat flux cooling,” IEEE Transactions on Advanced
Packaging, vol. 28, pp. 216-230, 2005.
[19] V. Benetics, A. Shooshtari, and P. Foroughi, “A source-integrated micropump
for cooling of high heat flux electronics,” Nineteenth Annual IEEE
Semiconductor Thermal Measurement and Management Symposium, San Jose,
CA, March 11-13, pp. 236-241, 2003.
14
[20] C.K. Lee, A.J. Robinson, and C.Y. Ching, “Development of EHD Ion-Drag
Micropump for Microscale Electronics Cooling Systems,” 13
th
International
Workshop on Thermal Investigation of ICs and Systems, Budapest, Hungary,
September 17-19, pp. 48-53, 2008.
[21] H. Ma, B. Hou, J. Gao, and C.Lin, “Development of One-sided Actuating
Piezoelectric Micropump Combined with Cold Plate in a Laptop,” Twenty-
fourth Annual IEEE Semiconductor Thermal Measurement and Management
Symposium, San Jose, CA, March 16-20, pp. 124-131, 2007.
[22] H. Lintel, F. Pol, and S. Bouwstra, “A piezoelectric micropump based on
micromachining of silicon,” Sensors and Actuators, vol. 15, pp.153-167, 1988.
[23] M. Koch, A.G.R. Evans, and A. Brunnschweiler, “The dynamic micropump
driven with a screen printed PZT actuator,” Journal of Micromechanics and
Microengineering, vol. 8, pp. 119-122, 1999.
[24] C. Qui, Q. Zhao, H. Zhang, W. Qu, H. Liu, and M. Cao, “A Valve-less PZT
Micropump with Isosceles Triangles Cross-section Diffuser Elements,” 1
st
IEEE
International Conference on Nano/Micro Engineered and Molecular Systems,
Zhuhai, China, January 18-21, pp. 200-203, 2006.
[25] S. Guo, J. Wang, and Q. Pan, “Solenoid Actuator-based Novel Type of
Micropump” IEEE International Conference on Robotics and Biomimetics,
Kunming, China, December 17-20, pp.1281-1286, 2006.
[26] M. Shen and M.A.M Gijs, “High-performance magnetic active-valve
micropump,” Transducers ’09, International Solid-State Sensors, Actuators and
Microsystems Conference, Denver, CO, June 21-25, pp. 1281-1286, 2009.
[27] S. Guo and J. Oohira, “A novel type of micropump using SMA actuator for
microflow application,” Proceedings of 2003 International Symposium on
Micromechatronics and Human Science, Nagoya, Japan, October 19-22, pp. 45-
50, 2003.
[28] R.B. Merrifield, “Solid Phase Peptide Synthesis. I. The Synthesis of a
Tetrapeptide,” Journal of the American Chemical Society, vol. 85, pp. 2149-
2154, 1963.
[29] A.R. Mitchell, S.B.H. Kent, M. Engelhard, and R.B. Merrifield, “A New
Synthetic Route to tert-Butyloxycarbonylaminoacyl-4-(oxymethyl)phenylacet-
amidomethyl-resin, and Improved Support for Solid-Phase Peptide Synthesis,”
Journal of Organic Chemistry, vol. 43, pp. 2845-2852, 1978.
15
[30] S.S. Wang, “p-Alkoxybenzyl Alcohol Resin and p-Alkoxybenzyloxycarbonyl-
hydrazide Resin for Solid Phase Synthesis of Protected Peptide Fragments,”
Journal of American Chemical Society, vol. 95, pp. 1328-1333, 1973.
[31] G.R. Matsueda and J.M. Stewart, “A p-Methylbenzhydrylamine Resin for
Improved Solid-Phase Synthesis of Peptide Amides,” Peptides, vol. 2, pp. 45-50,
1981.
[32] P. Sieber, “A new acid-labile anchor group for the solid-phase synthesis of C-
terminal peptide amides by the Fmoc method,” Tetrahedron Letters, vol. 28, pp.
2107-2111, 1987.
[33] R. Frank, “Spot-Synthesis: An Easy Technique for the Positionally Addressable,
Parallel Chemical Synthesis on a Membrane Support,” Tetrahedron, vol. 48, pp.
9217-9232, 1992.
[34] R. Frank, “The SPOT-synthesis technique. Synthetic peptide arrays on
membrane supports—principles and applications,” Journal of Immunological
Methods, vol. 267, pp. 13-26, 2002.
[35] N. Heine, T. Ast, J. Schneider-Mergener, U. Reineke, L. Germeroth, and H.
Wenschuh, “Synthesis and screening of peptoid arrays on cellulose membranes,”
Tetrahedron, vol. 59, pp. 9919-9930, 2003.
[36] X. Espanel, R.H. Huijsduijnen, “Applying the SPOT peptide synthesis
procedure to the study of protein tyrosine phosphatase substrate specificity:
probing for the heavenly match in vitro,” Methods, vol. 35, pp. 64-72, 2005.
[37] H.E. Blackwell, “Hitting the SPOT: small-molecule macroarrays advance
combinatorial synthesis,” Current Opinion in Chemical Biology, vol. 10, pp.
203-212, 2006.
[38] K. Hilpert, D.FH. Winkler, R.EW. Hancock, “Peptide arrays on cellulose
support: SPOT synthesis, a time and cost efficient method for synthesis of large
numbers of peptides in a parallel and addressable fashion” Nature Protocols,
vol. 2, pp. 1333-1349, 2007.
[39] D.H. Kim, D.S. Shin, and Y.S. Lee, “Spot arrays on modified glass surface for
efficient SPOT synthesis and on-chip bioassay of peptides,” Journal of Peptide
Science, vol. 13, pp. 625-633, 2007.
[40] S.A. Elrod, B. Hadimioglu, B.T. Khuri-Yakub, E.G. Rawson, E. Richley, N.N.
Mansour, and T.S. Lundgren, “Nozzleless droplet formation with focused
16
acoustic beams” Journal of Applied Physics, vol. 69, pp. 3341-3347, 1989.
[41] J.W. Kwon, H. Yu, Q. Zhou, and E.S. Kim, “Directional Ejection of Liquid
Droplets by Sectored Self-Focusing Acoustic Transducer Built on ZnO and
PZT,” Journal of Micromechanics and Microengineering, vol. 16, pp. 2697-
2704, 2006.
[42] J.W. Kwon, S. Kamal-Bahl, and E.S. Kim, “In-situ DNA Synthesis on Glass
Substrate for Microarray Fabrication Using Self-Focusing Acoustic Transducer,”
IEEE Transactions on Automation Science and Engineering, vol. 3, pp. 152-158,
2006.
[43] C.Y. Lee, S. Kamal-Bahl, H. Yu, J.W. Kwon, and E.S. Kim, “On-Demand
DNA Synthesis on Solid Surface by Four Directional Ejectors on a Chip,”
IEEE/ASME Journal of Microelectromechanical Systems, vol. 16, pp. 1130-
1139, 2007.
[44] D.G. Grier, “A revolution in optical manipulation,” Nature, vol. 429, pp. 810-
816, 2003.
[45] Y. Roichman, and D.G. Grier, “Projecting extended optical traps with shape-
phase holography,” Optics Letters, vol. 31, pp. 1675-1677, 2006.
[46] B. Landenberger, H. Höfemann, S. Wadle, and A. Rohrbach, “Microfluidic
sorting of arbitrary cells with dynamic optical tweezers,” Lab on a Chip, vol. 12,
pp. 3177-3183, 2012.
[47] M.B. Rasmussen, L.B. Oddershede, and H. Siegumfeldt, “Optical Tweezers
Cause Physiological Damage to Escherichia coli and Listeria Bacteria,” Applied
and Environmental Microbiology, vol. 74, pp. 2441-2446, 2008.
[48] K.C. Neuman, E.H. Chadd, G.F. Liou, K. Bergman, and S.M. Block,
“Characterization of Photodamage to Escherichia coli in Optical Traps,”
Biophysical Journal, vol. 77, pp. 2856-2863, 1999.
[49] J. Wu, “Acoustical Tweezers,” Journal of the Acoustical Society of America,
vol. 89, pp. 2140-2143, 1991.
[50] J. Lee, S.Y. Teh, A. Lee, H.H. Kim, C. Lee, and K.K. Shung, “Single beam
acoustic trapping,” Applied Physics Letter, vol. 95, pp. 073701-1-073701-3,
2009.
[51] P.L. Marston, “Axial radiation force of a Bessel beam on a sphere and direction
reversal of the force,” Journal of the Acoustical Society of America, vol. 120, pp.
17
3518-3524, 2006.
[52] G. Milne, G.D.M. Jeffries, and D.T. Chiu, “Tunable generation of Bessel beams
with a fluidic axicon,” Applied Physics Letters, vol. 92, pp. 261101-1-261101-3,
2008.
[53] C. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing
Air-Reflectors,” IEEE International Micro Electro Mechanical Systems
Conference, Istanbul, Turkey, January 22-26, pp. 170-173, 2006.
[54] E. Riera, J.A. Gallego-Juarez, and T.J. Mason, “Airborne ultrasound for the
precipitation of smokes and powders and the destruction of foams,” Ultrasonics
Sonochemistry, vol. 13, pp. 107-116, 2006.
18
Chapter 2
Hi-Fidelity Microspeaker on PZT Bimorph Diaphragm
This chapter describes piezoelectric MEMS microspeakers built on a PZT
bimorph diaphragm. In a four-edge-clamped diaphragm, a transducer with a bimorph
diaphragm is able to generate larger displacement compared to a transducer with a
unimorph diagram for a given driving voltage. For a given actuation voltage, a
microspeaker with top and bottom patterned electrodes showed much higher acoustic
pressure output than a microspeaker with a patterned top electrode and a un-
patterned bottom electrode in spite of possible misalignment between top and bottom
electrodes.
By optimizing top and bottom electrode patterns and having reasonable
alignment accuracy, the microspeaker having patterned top and bottom electrodes
showed flat acoustic pressure output of 122 dB SPL (± 3dB SPL) in the frequency
range of 1 kHz ~ 10 kHz. Since the main goal of microspeakers is generating large
sound output, and patterning electrodes does not increase device volume or require
higher actuation voltage level, a PZT bimorph diaphragm with top and bottom
patterned electrodes was chosen as the actuator for zero-velocity sensing application.
2.1 Transducer Design
Piezoelectric materials deform when electrical field is applied to the material,
and the deformation induces stress on the material. When design piezoelectric
19
transducers, placing electrodes on highly stressed area is very important to convert
the electric field into mechanical displacement effectively. Since the convergence
between electrical field and mechanical stress is bilateral, knowing stress distribution
of a four-edge-clamped diaphragm for a uniform acoustic pressure loading is
beneficial to deciding electrode patterns of an actuator. The normal stresses in x and
y directions (
x
and
y
) of a four-edge-clamped, isotropic diaphragm under a
uniform loading (
Z
) can be calculated using approximate equations (equation 2-1
and equation 2-2) with less than 10% error [1], where, a is diaphragm side and h is
diaphragm thickness.
[
(
)] (2-1)
(
)
(
)
[
(
)] (2-2)
(
)
The normalized contour plots of Equation 2-1 and Equation 2-2 are shown in
Figure 2.1 [1]. If we overlay the
xx
and
yy
plots, we can see the opposite type of
stress (i.e., tensile and compressive stress) exist at the same time at certain locations
(Figure 2.2a). These opposite-type stresses at a same area cancel out the actuation
effects, and decrease the bending displacement. Thus, the electrodes of the front side
20
and backside of the PZT bimorph diaphragm are placed so that the areas covered by
the electrodes have a same type of stress in both x and y directions (Figure 2.2b).
Figure 2.1 Contour plots of the normal stresses (a)
xx
and (b)
yy
Figure 2.2 Contour plots of (a) the area where the normal stresses
xx
and
yy
have
opposite sign and of (b) the summed normalized stress (
xx
+
yy
) for those areas
where
xx
and
yy
have same sign.
21
When a bimorph diaphragm deforms, the stress direction on upper layer is
opposite to the stress direction on lower layer. For instance, if we assume that the
bimorph diaphragm bends upward for a uniform acoustic loading, tensile stress is
induced at the center on upper layer while compressive stress is induced at the same
location on lower layer (Figure 2.3). The stress directions at the four clamped edges
are all the same to each other and opposite to the stress direction at the center.
2.2 Fabrication
2.2.1 Bimorph Formation
Two PZT sheets were glued back-to-back using low viscosity glue (resin
epoxy) to form a PZT bimorph diaphragm so that polarization directions of each PZT
sheet point to each other. It is important to reduce the thickness of the low viscosity
glue layer between two PZT sheets, in order to maximize the electrical field inside
the PZT bimorph diaphragm because the glue layer takes some of the applied voltage
away. To calculate the electric field loss in the low viscosity resin epoxy glue, we
measured total capacitance (C
total
) of the bimorph diaphragm which is equivalent to
Figure 2.3 The cross-sectional view of stress distributions on a four-edge-clamped
bimorph diaphragm when a uniform loading is applied to the diaphragm.
22
the total capacitance of three capacitors (C
PZT-top
, C
glue
, and C
PZT-bottom
) connected in
series. The total capacitance is calculated using following equation.
(2-3)
From equation 2-3, we can calculate the relative dielectric constant of the low
viscosity resin epoxy using equation 2-4.
(2-4)
The calculated relative dielectric constant of the low viscosity epoxy resin was 63
which is about 20 times less than the dielectric constant of the PZT sheet. In other
words, the electric field loss per unit thickness in the low viscosity epoxy is 20 times
larger than that in the PZT sheet. Therefore, minimizing the glue layer thickness was
essential to improve actuation efficiency.
The low viscosity glue was spin-coated onto the PZT sheet at 6,000 RPM
rotational speed on a spinner to have very thin glue layer. Due to the high surface
roughness of the PZT sheet (typical thickness variation is as large as ± 5 μm), the
glue layer thickness was not uniform along the boundary between two PZT sheets
and the thickness went up to 4.24 μm at certain area. Non-uniform glue thickness can
be interpreted that mechanical properties of the diaphragm such as stiffness,
elasticity, and stringiness of mechanical binding between two PZT sheets are not
uniform. The actuation efficiency can be decreased because of the non-uniform
mechanical properties inducing in-plane direction resonance, which usually observed
23
in higher harmonic components in frequency-to-displacement measurement.
Reducing glue thickness is always desirable to minimize electric field loss in the glue
layer.
Mechanical polishing using diamond lapping films was performed on PZT
surface. The PZT surface lapping started with lapping film with grain size of 5 μm,
and the grain size was reduced to 1 μm and 0.1 μm. DI water was used as lubricant.
With the mechanical polishing of the PZT surface and the spin-coating of the low
viscosity resin epoxy at 6,000 RPM rotational speed, we were able to reduce the glue
layer thickness to about 1 μm with thickness variation less than 0.5 μm (Figure 2.4).
2.2.2 Process Flow for the PZT Bimorph Microspeakers
The microspeaker was formed by gluing two 127 μm thick PZT sheets and
attaching them onto a micromachined silicon substrate as shown in Figure 2.5. A
pre-deposited nickel layer was completely etched out on the same side of two PZT
sheets, and the etched surface was mechanically polished with diamond lapping
films. Low viscosity glue with viscosity less than 150 cps was spin-coated onto one
Figure 2.4 SEM photos of the cross-section of the fabricated PZT bimorph
diaphragm with (a) and without (b) the mechanical polishing.
24
PZT sheet at 6,000 RPM rotational speed, and then two PZT sheets were combined
back-to-back. The low viscosity glue was cured at around 70 °C to reduce curing
time and decrease the viscosity further. After the low viscosity glue is completely
cured, 0.3 μm thick nickel layer was added onto both sides of the PZT bimorph
diaphragm. The top and bottom electrodes were patterned with two sides of the
square diaphragm as alignment marks. The patterned diaphragm was aligned and
temporarily fixed to a micromachined silicon wafer in a mask aligner, and
completely glued with low viscosity glue through a wicking process driven by
capillary force [2]. The bulk micromachined silicon top-cover provides the clamped
boundary condition for the PZT bimorph diaphragm.
Figure 2.5 Brief fabrication steps for a PZT-based bimorph microspeaker
25
2.3 Experiment Setup
The microspeakers were packaged on a dual-in-line package (DIP) and the
electrodes were connected using silver paste shown in Figure 2.6. The displacements
were measured at the center of the 8 x 8 mm
2
diaphragm to characterize the
mechanical responses of the fabricated microspeakers (the microspeaker with
patterned top and bottom electrodes and the microspeaker with patterned top
electrode and un-patterned bottom electrode) with an Optodyne
TM
Laser Doppler
Displacement Meter (LDDM) which has 6.3 nm resolutions. Due to the rough PZT
surface, nickel electrode does not reflect the focused laser beam efficiently, and we
attached 1.5 x 1.5 mm
2
aluminum foils with double sided tape to enhance the laser
beam reflection for the LDDM measurements. The displacements were measured
with the microspeakers driven by 190V
peak-to-peak
sinusoidal signals from 50 Hz to 20
kHz, in a setup shown in Figure 2.7.
Figure 2.6 Photos of the fabricated microspeaker taken from (a) the backside and
from (b) the front side.
26
Figure 2.7 Testing setup for the displacement measurement of the acoustic
transducer
Figure 2.8 Testing setup for the sound pressure output measurement of the acoustic
transducer
27
The sound pressure level (SPL) was measured from 5 mm away from the
diaphragm with a LinearX
TM
M31 calibrated microphone with the microspeaker
driven by 190V
peak-to-peak
sinusoidal signals from 50 Hz to 20 kHz. The microspeaker
was encapsulated with a cylindrical plastic package having about 0.7 cc volume, and
the SPL of the microspeaker was measured in a setup shown in Figure 2.8. The
captured voltage level from the M31 calibrated microphone was transferred to a
computer and translated into SPL. The frequency was swept from 50 Hz to 20 kHz
with 800 equidistant data points in log scale.
2.4 Measurement Results
2.4.1 Mechanical Response of Microspeakers
With PZT bimorph microspeakers being driven with 190V
peak-to-peak
sinusoidal
signals, the bending displacement of the microspeaker, which has patterned top
electrode and un-patterned bottom electrode, was measured at the center of the
diaphragm with LDDM. The aluminum foil attached at the measuring point has
thickness of 16 μm and density of 2.1 g/cm
3
while the PZT bimorph diaphragm has
thickness of 256 μm and density of 7.8 g/cm
3
. The mass of PZT bimorph diaphragm
for 1.5 x 1.5 mm
2
is 600 times larger than that of the aluminum foil. In case of
thickness mode resonance, the frequency shift due to a mass loading can be
estimated by Sauerbrey equation [3], and the estimated frequency shift at 20 kHz due
to the 600 times mass difference is about 33 Hz. However, the PZT bimorph
microspeaker relies on bending moments and is less influenced by the mass loading
28
effect. Therefore, we can neglect the mass loading by the aluminum foil. The
displacements were measured to have 0.2 ~ 0.3 μm uniform displacements from 50
Hz to 10 kHz and also to have 1.6 μm displacements at the fundamental resonance
frequency of 16.2 kHz as shown in Figure 2.9 [4].
With same electrical driving condition, the mechanical response of the
microspeaker with patterned top and bottom electrodes was characterized to have 0.9
~ 1.3 μm from 100 Hz to 9 kHz. The maximum displacement was measured to be
7.25 μm at its fundamental resonance frequency of 14.3 kHz (Figure 2.10) [5]. With
a same PZT bimorph diaphragm, patterning bottom electrode improved the bending
actuation efficiency so that the displacements of the diaphragm increased by the
factor of four on both flat response region and fundamental resonance frequency.
The microspeaker with patterned top and bottom electrodes experiences
approximately two times larger electric field, when driven with same voltage level,
in thickness direction of the PZT bimorph diaphragm compared to the microspeaker
with patterned top electrode and un-patterned bottom electrode. The un-patterned
bottom electrode floats supplying intermediate voltage level between ground and
applied voltage. Employing un-patterned bottom electrode eliminates the needs of
fine alignment between top and bottom electrodes supplying electrical polarization
with opposite signs along the thickness direction to PZT bimorph diaphragm.
However, the un-patterned bottom electrode sacrifices the electrical field intensity by
factor of two. The two times difference of electric field intensity is translated into
four times difference of diaphragm displacements.
29
Figure 2.9 Measured displacements of the microspeaker with patterned top
electrode and un-patterned bottom electrode
Figure 2.10 Measured displacements of the microspeaker with patterned top and
bottom electrode
30
2.4.2 Acoustic Response of Microspeakers
The microspeaker with un-patterned bottom electrode produced sound pressure
level (SPL) of 103.6 ~ 110 dB between 1.3 kHz and 12 kHz and 125 dB at 16.2 kHz
as shown Figure 2.11 [4]. A cylindrical package which has 0.7 cc inside volume
encapsulated the microspeaker to enhance the low frequency response. The long
wavelength of acoustic waves in low frequencies reduces sound pressure output
because the acoustic energy density is proportional to the frequency of acoustic wave.
Since the detection area of the M31 calibrated microphone is fixed, lower acoustic
energy density is translated into lower sound pressure pickup at the microphone.
Encapsulating the microspeaker can enhance the sound pressure signals in low
frequencies by converting acoustic waves proceeding lateral direction to acoustic
waves proceeding toward calibrated microphone.
The microspeaker with top and bottom patterned electrodes was produced
sound pressure level (SPL) of 118 ~ 125 dB between 1 kHz and 10 kHz and 133 dB
at 14.3 kHz as shown in Figure 2.12 [5]. A cylindrical package with 0.7 cc inside
volume was used to encapsulate the microspeaker either. Due to the large
displacement of 1 μm between 100 Hz and 9 kHz, the microspeaker produces
uniform and large sound pressure level between 1 kHz and 10 kHz. Patterning both
top and bottom electrodes enhanced the sound pressure output by 15 dB SPL
(corresponding to the factor of 5.6 difference in magnitude) between 1 kHz to 10
kHz and 8 dB SPL (corresponding to the factor of 2.5 difference in magnitude) at the
fundamental resonance peak.
31
Figure 2.11 Sound pressure outputs measured at 5 mm away from the microspeaker
having patterned top electrode and un-patterned bottom electrode with a 0.7 cc
cylindrical encapsulating package
Figure 2.12 Sound output pressure measured at 5 mm away from the microspeaker
having patterned top and bottom electrodes with a 0.7 cc cylindrical encapsulating
package
32
2.4.3 Power Consumption Measurement of Microspeakers
In general, a piezoelectric transducer consumes less power than an
electromagnetic transducer because a piezoelectric transducer charges and discharges
its input capacitance rather than to flow significant amount of current like an
electromagnetic transducer. A power measurement setup shown in Figure 2.13 was
established to compare the power consumption of the PZT bimorph microspeaker
with top and bottom patterned electrodes and the power consumption of a
commercial electromagnetic cell-phone microspeaker, which was detached from a
commercial cell-phone. A 1Ω test resistor was connected in series with amplified
electrical signal which drives the microspeaker. The voltage drop across the test
resistor was measured with oscilloscope, and the captured wave forms were sent to a
computer to be converted into current value. The voltage wave forms which driven
to the microspeaker were also captured and sent to the computer through the other
signal channel.
Figure 2.13 Power measurement setup for microspeakers
33
The multiplication of measured voltage and current amplitudes represents a
complex power. The real power consumption can be obtained by multiplying the
cosine of the phase difference between the measured voltage and current to the
measured complex power. The calculated the real power consumptions of the
commercial electromagnetic microspeaker and the fabricated PZT bimorph
microspeaker were plotted in Figure 2.14. The real power consumption of the PZT
bimorph microspeaker was less than that of the commercial electromagnetic
microspeaker by an order of magnitude in frequency range from 100 Hz to 2 kHz,
and the power consumption increased as the frequency approaches to its resonance
frequency. Considering that commercial cell-phones cut off audio signals higher than
4 kHz, the fabricated piezoelectric microspeaker can reduce power consumption by
an order of magnitude if it replaces the commercial electromagnetic speakers.
Figure 2.14 Measured real power consumption of the PZT bimorph microspeaker
and a commercial cell-phone speaker
34
2.5 Summary
In spite of the great piezoelectric constant of a PZT, bimorph structure of a
PZT diaphragm has not been tried due to the difficulty of fabrication. To make
bimorph structure with sol-gel PZT, it is essential to have metal layer between the
two sol-gel PZT layers to polarize the upper and lower diaphragm in opposite
direction by applying opposite direction of electric field. Controversially, pacing
conducting layer between two layers of a bimorph diaphragm eliminates the
advantage of bimorph structure in generating bending momentum.
Acoustic transducers built on PZT bimorph diaphragm with very thin glue
layer have been designed and fabricated. The microspeaker with top and bottom
patterned electrodes was shown to be much more efficient than the microspeaker
with patterned top electrode and un-patterned bottom electrode. The twice stronger
electric field in the PZT bimorph diaphragm microspeakers with both-electrodes-
patterned resulted in 7.25 μm displacement, which is four times larger than that of
the other microspeaker, at the center of the diaphragm with same level of electrical
actuation voltage. The sound output was also turned out 5.6 times larger between 1
kHz to 10 kHz and 2.5 times larger at resonance peak in the microspeaker with both-
electrodes-patterned than the other.
35
Chapter 2 References
[1] E.S. Kim, “Integrated microphone with CMOS circuits on a single chip”, Ph.D
dissertation, University of California, Berkeley, 1990.
[2] J.W. Kim, S. Kamal-Bahl and E.S. Kim, “Film Transfer and Bonding
Technique to Cover Lab on a Chip,” Transducers ’05, IEEE International
Conference on Solid-State Sensors and Actuators, (Seoul, Korea), June 5-9, pp.
940-943, 2005.
[3] V.G. Sauerbrey, “Verwendung von Schwingquarzen zur Wägung dünner
Schichten und zur Mikrowägung”, Zeitschrift für Physik, vol. 155, pp. 206-222,
1959.
[4] Y. Choe, S.J. Chen, and E.S. Kim, “Bimorph Diaphragm Formed by Two PZT
Sheets on Micromachined Silicon for Sound Generation,” 6
th
International
Conference and Exhibition on Device Packaging, Scottdale/Fountain Hills, AZ,
March 9-11, pp. 269-270, 2010.
[5] Y. Choe, S.J. Chen, and E.S. Kim, “High Fidelity Loud Microspeaker Based on
PZT Bimorph Diaphragm,” Technical Proceeding of the 2010 NSTI
Nanotechnology Conference & Expo, Anaheim, CA, June 21-25, vol. 2, pp.
316-319, 2010.
36
Chapter 3
Valveless Micropump Driven by Acoustic Streaming
This chapter describes valveless micropumps built on a 260 μm-thick PZT with
20 μm-thick parylene acoustic Fresnel lenses with air cavities [1]. The micropumps
produce in-plane body forces through an acoustic streaming effect of a high intensity
acoustic beam that is generated by constructive interference of acoustic waves. The
fabricated micropump with U-shape acoustic lens arrangement was shown to move
microspheres, which have diameter of 70 ~ 90 μm and density of 0.99 g/cm
3
, on
water surface, and driven by 160 V
peak-to-peak
pulsed sinusoidal waves. Both the
acoustic Fresnel lens and the top electrode were patterned in a pie-shape with its
apex angle of 90° to form asymmetric acoustic pressure distribution at the focal
plane of the acoustic Fresnel lenses in order to push water in one direction. The
driven microspheres by the micropump with U-shaped acoustic lens arrangement
formed U-shape streaming even without any fluidic channel according to the serial
connection of the pie-shaped lenses and the top electrodes. A novel experiment setup
was devised to measure pumping rate in volume in circumstances where very low
back pressure exists. The measurement setup inherently underestimates the pumping
rate, and the micropump with an acrylic reflector and straight-lined acoustic Fresnel
lens arrangement showed very high pumping rate of 9.5 mL/min.
37
3.1 Acoustic Fresnel Lens Design
An acoustic Fresnel lens is a planar lens that uses constructive and destructive
interference of acoustic waves to focus the waves, and is typically built on annular
rings forming half-wave-band sources [2] that produce enhanced out-of-plane force
but balanced in-plane force at its focal point. Asymmetric in-plane particle
displacement (or an in-plane force) can be generated by sectoring the annular rings
and electrodes into a pie-shape as shown in Figure 3.1 [3]. For pumping liquid, we
explored two types of Fresnel lens, and ran MATLAB simulations (Figure 3.1) to
estimate the in-plane particle displacement generated by a single-pie-shaped Fresnel
lens with 90 ° sector angle. Both designs gave similar level of the in-plane particle
displacement at the focal plane. However, 50% more number of lenses can be packed
in a given area (i.e. allowing a larger net active area) with the lens with only 5 outer
rings (Figure 3.1a).
Figure 3.1 MATLAB simulation results showing in-plane particle displacements on
the focal plane of the Fresnel lens composed of (a) 5 outer rings and (b) all possible
rings. The yellow pie-shaped annular lines represent Fresnel lens shape.
38
A number of the sectored Fresnel lenses can be arranged in series to form a
liquid pump or transporter [4]. Both U-shaped arrangement (Figure 3.2) and straight-
lined arrangement of the Fresnel lenses can produce unidirectional liquid flow. In
case of the micropump based on U-shaped arrangement of the sectored lenses, the
lenses were arranged in such a way that the body force due to the focused acoustic
beam points along the direction tangential to the curved line of the U-shaped channel
to minimize the friction loss between water and fluidic channel (Figure 3.2). In case
of the micropump based on straight inline particle movements, the Fresnel lensed
were arranged on a straight line with minimum spacing between the lenses for
maximum lens density in a given area. Both sectored Fresnel lens utilizing all
possible rings with and sectored Fresnel lens utilizing only 5 outer rings were
designed and fabricated for comparison.
Figure 3.2 Arrangement of the sectored Fresnel lenses in U-shaped fluidic channel.
The body force due to the focused acoustic beam is designed to be directed along the
line tangential to the curved shape of the U-shaped channel.
39
3.2 Fabrication
The micropumps were fabricated in the following steps. After depositing 0.3
μm-thick additional nickel layer on top of the 0.1 μm-thick pre-deposited nickel
layers to reduce series resistance of electrodes, the top and bottom electrodes
sandwiching 260 μm-thick PSI-5H4E PZT sheet were patterned, using the two sides
of a rectangular PZT substrate as alignment lines (Figure 3.3a), and then sacrificial
photoresist layer was spin-coated and patterned to form Fresnel lens pattern (Figure
3.3b). After 3.5 μm-thick parylene D layer was deposited on the photoresist as a
structure layer (Figure 3.3c), the sacrificial photoresist layer was removed through
release holes (30 μm in diameter) in acetone for 24 hours (Figure 3.3d). Additional 7
μm-thick parylene D layer was deposited to seal the release holes (Figure 3.3e), and
patterned to open up contact holes for the top and bottom electrodes access. After
wires are connected to the electrode pads, 10 μm-thick parylene D layer was
deposited for electrical insulation between top and bottom electrodes.
Figure 3.3 Brief fabrication steps of the micropumps
40
Figure 3.4 Photos of the fabricated micropumps with U-shaped channels: (a) the
micropump built on the sectored Fresnel lens shown in Figure 3-1b and (b) the
micropump built on the Fresnel lens utilizing only 5 outer rings of the Fresnel lens
shown in Figure 3-1a
Figure 3.5 Photos of the fabricated micropumps having a straight-lined channel (a)
without and (b) with an acrylic reflector
41
3.3 Measurement Setups
Various operating conditions were used to find the optimal operation
conditions for the fabricated micropumps, since there are many parameters that
affect the pump performance such as pulse repetition frequency (PRF) and the
distance between the focal plane of the Fresnel lenses and the liquid surface.
3.3.1 Micropumps with U-shaped Channel
The fabricated micropumps were submerged in DI water with the distance
between the water surface and the micropump being about 4.5 mm, and actuated by
pulsed 8.6 MHz sinusoidal signals with 20 kHz PRF and 30 % duty cycle. The
pulsed sinusoidal signals were amplified by a power amplifier to have voltage
amplitude of 160 V
peak-to-peak
. To facilitate the observation of the liquid pumping,
microspheres having a density of 0.99 g/cm
3
were let to float on the water surface. A
video camera captured the movements of the microspheres from the top of the
micropump, and the drift velocity of the microspheres was calculated later by
analyzing the captured video (Figure 3.6).
Figure 3.6 Measurement setup for the micropump with U-shaped channel
42
3.3.2 Micropumps with Straight-lined Channel
The fabricated micropump with a straight inline channel also were
characterized in various test conditions such as PRF being varied from 40 Hz to 300
Hz, the micropump (that was designed to have a focal length of about 5 mm) being
submerged 3 mm, 5 mm, and 7 mm below the water surface.
Since typical description about pumping ability is written in volume transfer
rate per unit time, it is better to have an experiment setup which measures the
pumping rate in volume transfer rate in unit time rather than particle drift velocity on
the liquid surface. Measuring accurate pumping speed of the fabricated micropump
requires a special setup, as any backpressure affects the pumping speed. We devised
a novel measurement method, which inherently underestimates the pumping rate
though, to measure the pumping rate of the micropump with straight line channel
(Figure 3.7). The micropump was located between a water reservoir (Figure 3.7a)
and a water drain tube (Figure 3.7c) that was placed at the same level as the reservoir
and the micropump to minimize any hydrostatic pressure difference between the
reservoir and drain. First, the water flow rate due to the hydrostatic pressure of the
water in the reservoir was measured without actuating the micropump. After
restoring the water level to the initial condition by pouring the collected water from
the drain back into the reservoir, we measured the flow rate again, this time the
micropump being actuated. The pumping rate of the micropump was obtained by
taking the difference between the two measured flow rates.
43
Figure 3.7 Photo from top view of the setup to measure pumping rate of the
micropump: (a) a reservoir for DI water, (b) an inline micropump, (c) a water drain
tube placed on the same level with the liquid channel to minimize the differential
pressure between the reservoir and the drain tube.
Figure 3.8 Photo from side view of the setup to measure pumping rate of the
micropump
44
3.4 Measurement Results
3.4.1 Micropump with U-shape Channel
The micropump with the full Fresnel lens pattern (Figure 3.4a) did not show
any consistent microsphere movement possibly due to relatively small active area
and inconsistent in-plane acoustic pressure. However, the micropump, which utilize
5 outer rings of Fresnel lens only (Figure 3.4b), strongly moved microspheres on
liquid surface.
From the captured video (Figure 3.9), a group of polystyrene particle moved
7.3 mm during 3 frames. Since the video camera of the measurement setup captures
30 frames each second, the equivalent time for 3 frames is about 100 msec.
Therefore, the particle drift speed can be calculated using following equation, and
turn out to be about 7.3 cm/sec.
(3-1)
Figure 3.9 Photos (from a video) showing that a group of polystyrene microspheres
were moved about 7.3 mm in 100 msec.
45
3.4.2 Micropump with Straight-lined Channel without a Reflector
The fabricated micropump with straight-lined channel was actuated with
various channel width (6 mm, 8mm, and 10 mm) to see the effects of different
channel width on pumping efficiency. The micropump having 1.5 mm focal length
of the Fresnel lenses and 90° sector angle for both Fresnel lens and electrode was
submerged 3 mm below from the water surface, and actuated with 8.544 MHz pulsed
sinusoidal signals with 160 PRF and 30% duty cycle. Under same electrical driving
condition, the micropump showed similar pumping speed and similar width of
polystyrene stream as shown in Table 3.1.
Since the in-plane particle displacement is formed by focused acoustic beam,
the most vigorous liquid stream is observed along the center line of Fresnel lens.
Though the water level is lower than the focal plane of the Fresnel lens, the width
water stream is narrow as about 2 mm due to the focusing effect of the Fresnel lens.
If the channel width is larger than that of water streaming, the effect of channel
width is almost negligible.
The drift velocity calculated from captured video images become inaccurate as
the drift velocity become higher because the video images capturing fast moving
particle get blurred unless the capture rate increased accordingly. Therefore, the
velocity differences in range of few millimeter/sec can be reasonably neglected, and
it is safe to say that the micropump with difference channel width showed same level
of pumping ability regardless to the channel width.
46
Table 3.1 Particle drift velocity of the micropump with straight-lined channel
Channel Width
Particle Drift Velocity
(mm/sec)
Width of Liquid Stream
(mm)
6 mm 80.1 2.4
8 mm 83.2 2.3
10 mm 81.2 2.0
Figure 3.10 Photos showing width of fast water stream according to different
channel width: (a) 6 mm, (b) 8 mm, and (c) 10 mm
47
The micropump with a straight-lined channel was also tested to see the relation
between various test conditions and particle drift velocity. The operating conditions,
in which the micropump produces unidirectional liquid flow, were recorded, and the
measured drift velocities versus PRF as a function of water height are plotted in
Figure 3.11. As can be seen in Figure 3.11, the pumping speed is strongly dependent
on the PRF and water level. Optimal PRF for the highest pumping speed depends on
the water height. The curves for the 3 mm water level and 7 mm water level showed
somewhat similar trends (unlike the one for the 5 mm water level), possibly because
the two cases have same distance away from the focal plane of the micropump albeit
one in positive direction and the other in negative.
Figure 3.11 Measured particle drift velocities vs. pulse repetition frequency (PRF) as
a function of the water height
48
As we can see in Figure 3.11, the curve for the 3 mm water level is positioned
higher than the curve for the 7 mm water level. Though both cases have same
distance to the focal plane, the curve for the 7 mm water level is lower because the
acoustic waves are attenuated as they proceed in the water, and have lower in-place
particle displacement after they travel 7 mm from the micropump (Figure 3.13).
When observed on the focal plane of the Fresnel lens, acoustic waves are
focused to the focal spot of the acoustic lens having high in-plane and vertical
particle displacements. Because of the effective focusing of the Fresnel lens, the
particle displacement appears in relatively small area. When observed lower level
than the focal plane of the Fresnel lens, the acoustic waves have not been effectively
focused yet. In other words, the acoustic waves are on their way to the focal plane
foaming in-plane force in larger area but with lower level of acoustic intensity
(Figure 3.12).
Having in-place particle displacement in large area is more advantageous for
micropump application because the low acoustic intensity due to poor focusing can
be improved by supplying higher electrical power to the micropump. However, if we
supply higher electrical power to the micropump with the water level placed on the
focal plane of the Fresnel lens, the highly focused acoustic beam tends to eject or
splash droplets at the focal spots rather than to drift the water to one direction. In the
Figure 3.11, the plot with 3 mm water level showed higher particle drift velocity than
other plots even with same electrical power supply.
49
Figure 3.12 MATLAB simulation results showing in-place particle displacement
observed at 2 mm away from acoustic Fresnel lens which has focal length of 4 mm
Figure 3.13 MATLAB simulation results showing in-place particle displacement
observed at 6 mm away from acoustic Fresnel lens which has focal length of 4 mm
50
3.4.3 Micropump with Straight-lined Channel with a Reflector
Since Fresnel lens uses interference of acoustic waves, there could be
undesirable interference between the acoustic waves passing through Fresnel lenses
that are placed side by side. To resolve this issue, a laser-machined 2 mm-high
acrylic acoustic wave reflector was placed to eliminate the undesirable interaction
between adjacent Fresnel lenses and thus to improve the pumping efficiency.
Therefore, the waves that would have interfered with the adjacent areas are reflected
from the acrylic reflector that is shorter than the liquid height (>3mm) as shown in
Figure 14. When measured in identical setup with micropump without the reflector,
the micropump with the wave reflector not only showed 1.5 times higher particle
drift velocity but also showed more stable operation in wider range of PRF (Figure
3.14).
Figure 3.14 A conceptual diagram showing the working principle of the acrylic
reflector
51
In general, the maximum drift velocity is observed in the middle of the fluidic
channel. However, higher particle drift velocity is not necessarily translated better
pumping ability because the width of liquid stream also affects the actual pumping
rate. In such point of view, Figure 3.15 should be carefully interpreted. The curves
for 7 mm water level (drawn in red line) could be considered the curve that has most
wide range of stable operating conditions. However, considering both particle drift
velocity and the liquid stream width, we concluded that the optimal PRF to drive this
micropump is between 100 Hz to 106 Hz for all the three test conditions (3 mm, 5
mm, and 7 mm water level). The liquid stream width was maximum around 103 Hz
of PRF, and became narrower as the PRF recedes from optimal PRF of 103 Hz for
all three kinds of water levels. In the experiment with 7 mm water level, therefore,
the actual liquid pumping rate decreased despite the slight increase of particle drift
velocity on the liquid surface as the PRF increased from 103 Hz to 141 Hz because
of the liquid stream width being narrowed.
The pumping rate was also tested with different applied power level. The PRF
was fixed to its optimal value of 103 Hz when the micropump with an acrylic
reflector was in operation. As shown in Figure 3.16, the measured particle drift
velocity was also increased from about 40 mm/sec to 140 mm/sec when the applied
power to the micropump increased from 9.4 W to 34.8 W. Considering the ratios of
3.7(= 34.8/9.4) and 3.5(=140/40) between maximum and minimum applied powers
and particle drift velocities respectively, the particle drift velocity is linearly
proportional to the applied power.
52
Figure 3.15 Plot of pulse repetition frequency (PRF) vs. particle drift velocity of the
micropump with an acrylic reflector measured with different distances between the
micropump and the liquid surface.
Figure 3.16 Plot of applied power vs. particle drift velocity of the micropump with
an acrylic reflector measured with different distances between the micropump and
the liquid surface.
53
To measure volume transfer rate of the micropump with straight-lined channel
and a reflector, the measurement setup shown in Figure 3.7 and Figure 3.8 were used.
The water flow rate without micropump actuation was measured in volume for 30
seconds. After restoring the collected water from the drain back into the reservoir,
the water flow rate was measured again with micropump actuation at this time. By
taking the water flow rate difference between these two cases, the water pumping
rate of the fabricated micropump in volume was calculated to be 159 μL/sec (or 9.5
mL/min).
3.5 Summary
This chapter describes novel valveless micropumps with various Fresnel lens
designs for U-shaped and straight-lined liquid flow channels. The micropumps were
built on 10 μm-thick parylene Fresnel lens over a 260 μm-thick PZT sheet. The
micropump with U-shaped channel showed stirred polystyrene microparticles to
have particle drift velocity of 7.3 cm/sec, and the micropump with straight-lined
channel showed volume transfer rate of 9.5 mL/min. Both micropumps with straight-
lined and U-shaped channel used acoustic Fresnel lens which utilize only 5 outer
rings to maximize the number of acoustic lenses in a given design area.
Considering high performance micropumps, whether they are valveless or not,
reports pumping rate of several hundreds of micro liter per second, the fabricated
valveless micropump with reflector improved the pumping rate by an order of
magnitude.
54
Chapter 3 References
[1] C.Y. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing
Air-Reflectors,” the 19
th
IEEE International Conference on Micro Electro
Mechanical Systems, pp. 170-173, 2006
[2] D. Huang and E.S. Kim, “Micromachined Acoustic-Wave Liquid Ejector”,
IEEE/ASME Journal of Microelectromechanical Systems, vol. 10, pp. 442-449,
2001.
[3] V. Vivek, Y. Zeng, and E.S. Kim, “Novel Acoustic-Wave Micromixer,”
Thirteenth Annual International Conference on Micro Electro Mechanical
Systems, pp. 668-673, 2000.
[4] H. Yu, J.W. Kwon, and E.S. Kim, “Microfluidic Mixer and Transporter Based
on PZT Self-Focusing Acoustic Transducers,” IEEE/ASME Journal of
Microelectromechanical Systems, vol. 15, pp. 1015-1024, 2006.
55
Chapter 4
Peptide Synthesis on a Modified Glass Surface Using
Acoustic Droplet Ejector
This chapter describes a 9-mers-long peptide synthesis on a modified glass
surface using an acoustic droplet ejector with a silicon lens. To synthesize peptide
chain on a modified glass substrate, Spot synthesis protocol, which is described in
detail in Lee’s work, was followed [1]. In general, Spot synthesis is performed on
cellulose supports which increase the spot size up to centimeter range in diameter
because dispensed liquid spreads out in cellulous substrate and occupies large area
[2-4]. The modified glass surface is hydrophilic in amine-terminated spot, on which
amino acids are coupled, and hydrophobic outside of the coupling spot. Since the
hydrophobic glass surface prevents the dispensed liquid from spreading out, high
density of peptide array can be achieved by employing a surface modified glass slide
and an acoustic droplet ejector.
An acoustic micro-droplet ejector with a silicon lens replaced conventional
pipettes in dispensing pre-activated amino-acid solution on a surface modified glass.
The micro-droplet ejector ejects droplets with diameters of 80 μm that have
equivalent volume of 200 pL. Employing a droplet ejector as a liquid dispenser gives
us better volume control especially when very small amount of liquid is dispensed.
Since the acoustic ejector ejects a droplet per actuation pulse, precise volume control
56
can be achieved with a volume control resolution of 200 pL. A peptide consisting of
9-mers of glycine was successfully synthesized on a surface modified glass slide, and
the synthesis was optically confirmed at 532 nm using FITC staining.
4.1 Materials for Peptide Synthesis
Following chemicals were purchased from Sigma-Aldrich (USA): γ-APTS, 1-
hydroxybenzo-triazole (HOBt), acetic anhydride, benzo-triazol-1-yloxy-tris
(dimethylamino) phosphonium hexafluorophsphate (BOP), 2,2,4-trimethylpentane,
diisopropylethylamine (DIPEA), N-methylpyrrolidinone (NMP), diethylamine, and
Tween20. Fluorescein isothiocyanate (FITC) and Fmoc-amino acid (Gly) were
purchased AnaSpec (USA). 1H,1H,2H,2H-perfluorodecylmethyldichlorosilane was
purchased from AlfaAesar (USA). Chloroform, methyl alcohol, hexane, and acetone
were purchased from VWR (USA). The glass slides (25m x 75 mm) were purchased
from Corning Glass Works (USA).
4.2 Spot Protocol to Synthesize Peptides on a Glass Substrate
At first, a glass slide was cleaned with piranha solution (4 parts of H
2
O
2
and 1
parts of H
2
SO
4
) for 30 min. Then the glass slide was cleaned with acetone (3 times)
and chloroform (3 times), consecutively, by submerged in a beaker which is
containing about 20 mL of the chemical and shaken for 60 sec. To define active
spots, where amino acid will be attached later, AZ 5214 photoresist was spin-coated
onto the slide at 1,500 RPM rotational speed for 30 seconds, and exposed on contact
mask aligner with 80 mJ/cm
2
UV energy level. After developing the photoresist with
57
AZ 400K developer, the photoresist-patterned glass slide was submerged in 1% (v/v)
1H,1H,2H,2H-perfluoredecylmethyldichlorosilane in 2,2,4-trimethylpentane solution
for 15 min to form a hydrophobic perfluorinated layer. After washing the glass slide
5 times with n-hexane and 3 times with chloroform, the photoresist on the glass slide
was removed by acetone. The spot-arrayed glass was submerged in 5% (v/v) γ-APTS
in chloroform solution at 50 °C for 1 hour to silanize the active spots and to form
direct amine. After washed 5 times with chloroform, the glass slide was dried in
vacuum chamber for 1 hour before the first amino acid solution is dispensed onto the
surface modified glass slide.
Figure 4.1 Preparation of the amine-terminated glass surface
58
To couple amino acids on to the glass surface, pre-activated solution (6.7 mM,
200 nL) of fmoc-amino acid with 6.7 mM of BOP, HOBt, and DIPEA in NMP was
dispensed using an acoustic micro-droplet ejector onto the amine-terminated spots of
the glass and reacted for 60 min. The remaining un-coupled free amine groups were
capped using 10% (v/v) solution of acetic anhydride and pyridine (1:1) in NMP for
30 min at 25 °C. The fmoc-protection group was removed with 20% solution of
diethylamine in NMP for 30 min at 25 °C to liberate free-amine group. The liberated
free-amine group was coupled with next mer amino-acid or reacted with a 5.0 mM
solution of FITC in NMP for 2 hours at 25 °C for the optical measurement that
quantifies the number of the coupled amino acids.
Figure 4.2 Conceptual diagrams showing amino acid coupling and FITC attachment
procedures
59
4.3 Ejector Design and Fabrication
A vertical acoustic ejector with a lens with air reflector (LWAR) [5] was
designed using silicon lens. The ejector employs conventional Fresnel lens to
generate a focused acoustic beam on the liquid surface by focusing acoustic waves
generated from a 127 μm thick PZT transducer. As a structural material of the
acoustic Fresnel lens, parylene D film was replaced with a (100) silicon wafer with 3
inch diameter to simplify the device fabrication process and to make it easier to clean
ejectors by having smooth bottom surface in the shooting chamber.
An acoustic silicon lens was fabricated on a 400 μm thick silicon wafer. At
first, the wafer went through bulk-micromachining to make front-backside alignment
marks with LPCVD SiN
x
etch mask. After removing the SiN
x
with hot phosphoric
acid, the front side of the silicon wafer was dry-etched in deep reactive ion etcher
(DRIE) to form 20 μm deep trenches with Fresnel lens pattern. The backside was
also dry-etched to form ejection chamber, liquid transfer channel, and liquid
reservoirs. A PZT transducer was processed separately that starts with depositing 0.4
μm thick top and bottom nickel electrodes sandwiching a 127 μm thick PZT sheet.
Low viscosity glue with viscosity of 150 cps was spin-coated onto the PZT
transducer at 6,000 RPM rotational speed and combined with the silicon wafer. The
low viscosity glue was cured in elevated temperature of 80 °C for 24 hours. After
integration of the silicon lens and the PZT transducer, a bulk-micromachined silicon
top cover was added to maintain the liquid level constant. Then the complete ejector
was mounted on a dual-inline-package (DIP) socket.
60
Figure 4.3 Fabrication steps of an acoustic ejector with a silicon lens
Figure 4.4 Photos of fabricated device: (a) top view and (b) bottom view of the
packaged ejector. (c) an ejector without top cover.
61
4.4 Measurement Setup
For the device testing, the fabricated ejector was actuated by pulsed sinusoidal
signals with 60 Hz pulse repetition frequency (PRF) and 17.3 MHz driving
frequency. The pulsed signals were amplified with an RF power amplifier (75A250,
Amplifier Research) to have voltage amplitude of 160 V
peak-to-peak
and driven into the
ejector. A red light-emitting diode (LED) was used as a light source to
stroboscopically observe the ejection. Different moment of the ejection process was
observed by adjusting the delay between device actuation and LED illumination. A
charge-coupled device (CCD) camera (SONY SSC-DC54A) being attached at the
end of a long-working-range microscope was placed horizontally to record the
ejection process to a computer (Figure 4.5). The glass slide with free-amine on its
active spots was placed about 2 mm away from the ejector and aligned, so that the
ejected droplets of pre-activated amino-acid solution can ink the active spots on the
glass slide. Proteins of various lengths, consisting of Glycine, were synthesized on
different spots on the glass slide. After synthesizing the desired number of amino
acids on each active spot, we attached FITC fluorescent tag at the end of the
synthesized peptide. A fluorescent microscope (ef0046000a, Microscopes Inc.)
equipped with cooled CCD camera (TCC-3.3ICE-N, Tuscen) was used to optically
detect the signals coming from FITC at 532 nm to confirm the synthesis (Figure 4.6).
Between each amino acid coupling, the glass slide went through capping and fmoc-
deprotection steps with 2 times NMP washings and 1 time methyl alcohol washing
for each step.
62
Figure 4.5 Measurement setup of pre-activated amino acid solution ejection onto a
slide glass
Figure 4.6 Photo of the fluorescent microscope
63
4.5 Measurement Results
4.5.1 Optimum Coupling Time Measurement
Lab surroundings can affect experimental results of any bio-related
experiments. Therefore, finding optimal experiment condition in a given experiment
setup is very important even when well-known protocol is followed. A simple
experiment to find optimal amino-acid coupling time was performed on modified
glass surface. 200 nL of pre-activated amino-acid solution of glycine was dispensed
on amine-terminated glass surface by pipetting and coupled with various reaction
time. Since the light intensity coming from FITC labeled amino-acid represents the
amount of coupled amino-acids onto the glass surface, the coupling time, at which
the light intensity saturated, was investigated.
Figure 4.7 Optical detection of FITC using a fluorescent microscope
64
The coupling time was varied from 5 to 90 min (5, 10, 20, 30, 40, 60, and 90
min). An active spot was capped without amino acid coupling for negative control.
Three sets of samples were prepared in a same way. Since the absolute value of light
intensity varies from sample to sample because of experimental variation, the
measured light intensities were normalized to the light intensity of the brightest spot
(coupled with the pre-activated amino acid solution for 90 minutes) using following
equation and plotted in Figure 4.8.
(4-1)
Where, I
normalized
: Normalized light intensity
I
90min
: the light intensity of the spot with 90 min coupling time
I
capping only
: the light intensity of the spot treated capping solution only
I
n
: the light intensity of the spot with n min coupling time
Figure 4.8 Normalized light intensities measured from active spots with various
coupling times.
65
4.5.2 Ejecting Condition Calibration
The optimal driving condition for droplet ejection such as pulse width and
power level can be changed according to the type of ejected liquid. The 17.351 MHz
pulsed sinusoidal signals having 2 μsec pulse width with 60 Hz PRF was good
enough to eject droplets of DI water. However, the pulse width had to be increased to
6 μsec to eject droplets of pre-activated amino acid in NMP solution. Since NMP has
1.7 times higher viscosity (1.7×10
-3
Pa· s) than DI water (1.0×10
-3
Pa· s), longer pulse
width was required to eject NMP droplets. In other words, higher electric energy
should be applied to the ejector to eject droplets breaking strong molecular bonding
of NMP. Because of the strong molecular bonding, the vapor pressure of NMP (1.0
Torr at 40 °C) is also lower than that of DI water (58.3 Torr at 40 °C). The ejector
ejected droplets having 80 μm diameters as shown in Figure 4.9.
Figure 4.9 Photo of the pre-activated amino acid solution droplet ejection
66
4.5.3 Peptide Synthesis using Ejectors
A ladder structure of peptides (1-mer to 9-mers) consisting of glycine was
synthesized on 9 different spots on a surface modified glass slide. In a ladder
structure, each active spot in a column has 1-mer peptides length difference from
neighboring spots.
In the beginning, about 60 nL (equivalent with 300 droplets) of pre-activated
amino-acid solution was dispensed onto all 9 active spots of the ladder structure on
the glass slide using an acoustic ejector. The being dispensed spot was placed right
above the ejecting chamber opening (where droplets are vertically ejected) and about
2 mm away from the ejector. After finishing dispensing 60 nL of the solution on the
desired spot, the glass slide was moved so that the ejector can dispense the solution
onto the next active spot. Once the ejector dispenses the amino acid solution onto the
9 active spots of the ladder structure, the active spots were reacted with the amino
acid solution for 60 min.
Figure 4.10 Collected droplets of the pre-activated amino acid solution ejected by
the acoustic droplet ejector with a silicon lens. Each liquid bump has 60 nL of the
solution.
67
After washing the whole glass slide 2 times with 20 mL of NMP by shaking in
a beaker for 60 seconds and 1 time with methyl alcohol with the same method, the
glass slide was blow-dried with nitrogen and capped for 30 min by dispensing 200
nL of capping solution onto the amino-acid dispensed spots with a pipette. After
washing the capping solution in the same washing method above, fmoc-protection
groups of 8 active spots in the ladder structure were removed by dispensing 200 nL
of fmoc-deprotection solution to be coupled with next-mer amino acid. At this
moment, the 9
th
spot is not active anymore, and it has only 1-mer-long glycine
peptide to the end of the experiment. After washing fmoc-protection group
removing solution, the slide glass was moved onto the ejector to be coupled with
next mer amino acid. By repeating above processes, 9 spots in a column of spot array
have different lengths of peptide from 1-mer to 9-mers (Figure 4.11).
When 9
th
mer amino-acid coupling and capping were finished, fmoc-protection
group was removed in all the 9 spots of the ladder structure for FITC staining. After
FITC staining, the 9 spots were examined under fluorescent microscope at 532 nm to
measure the light intensities coming from the amino acids coupled onto the active
spots on the glass slide (Figure 4.12). Since the coupling efficiency is not 100 %, the
light intensity coming from the longer mers of peptide is lower than that coming
from the shorter mers of peptide. Comparing the measured light intensity of different
spots, we can calculate stepwise coupling efficiency or overall coupling efficiency to
synthesize certain number of mers because the light intensity represents the
population of coupled amino acid in the spot.
68
Figure 4.11 The ladder structure of peptide chains
Figure 4.12 FITC attached at the end of the peptide chain after synthesizing desired
number of peptides
69
Figure 4.13 Photos showing droplet’s position at (a) 150 μsec, (b) 450 μsec, (c) 700
μsec, and (d) 1000 μsec after the droplet is ejected from the ejection chamber.
Figure 4.14 Fluorescent images taken from the 9 spots in the peptide ladder structure
70
Figure 4.15 shows the monotonous decrease of light intensity as the length of
peptide increased from 1-mer to 9-mers if we exclude the light intensity measured
from 3-mers and 4-mers long peptide. The light intensities coming from these two
spots are much lower than expected value possibly because the quality of the free
amine group on those active spots were not as good as other spots. Since most spots
showed reasonable light intensity, the average stepwise coupling efficiency and
overall coupling efficiency of synthesizing a 9-mers-long glycine peptide were
calculated to be 96% and 70%, respectively.
Figure 4.15 Plot of measured light intensity
71
4.6 Summary
Protein microarray can be fabricated through either immobilization of pre-
synthesized protein or in-situ synthesis peptide synthesis. Spot technique is one of
the in-situ parallel peptide synthesis methods that is cost effective, easy to automate,
and has simpler experimental procedure. Cellulose membranes were one of the most
common substrates for Spot synthesis procedure, but replaced with surface modified
glass in this work to prevent the dispensed liquid from spreading out, and
consequently, to have higher peptide microarray density. Even with the glass
substrate, there still exists a limit to increase the peptide array density if the liquid
dispensing method is not improved from conventional pipetting.
An acoustic micro-droplet ejector with silicon lens was designed and fabricated
to supply advanced liquid dispensing technique with precise volume control
resolution of 200 pL. Combined with glass substrate, the acoustic ejector can
dramatically increase the peptide micro array density.
To prove the eligibility of utilizing a MEMS acoustic droplet ejector in Spot
peptide synthesis technique, optimal coupling condition was investigated. Also, 9-
mers-long peptide of glycine with a ladder structure was effectively synthesized on a
surface modified glass slide showing 96% stepwise coupling efficiency. Though
active spots with a diameter of 1 mm have been patterned in this experiment, the spot
size can be easily reduced further using photolithography technology.
72
Chapter 4 References
[1] D.H. Kim, D.S. Shin, and Y.S. Lee, “Spot array on modified glass surfaces for
efficient SPOT synthesis and on-chip bioassay of peptides,” Journal of Peptide
Science, vol. 13, pp. 625-633, 2007.
[2] M.D. Bowman, R.C. Jeske, and H.E. Blackwell, “Microwave-Accelerated
SPOT-Synthesis on Cellulose Supports,” Organic Letters, vol. 6, pp. 2019-2022,
2004.
[3] A. Kramer, U. Reineke, L. Dong, B. Hoffmann, U. Hoffmȕller, D. Winkler, R.
Volkmer-Engert, and J. Schneider-Mergener, “Spot synthesis: observations and
optimizations,” Journal of Peptide Research, vol. 54, pp. 1-9, 1999.
[4] R. Frank, “The SPOT-synthesis technique Synthetic peptide arrays on
membrane supports-principles and applications,” Journal of Immunological
Methods, vol. 267, pp. 13-26, 2002.
[5] C.Y. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing
Air-Reflectors,” the 19
th
IEEE International Conference on Micro Electro
Mechanical Systems, pp. 170-173, 2006.
73
Chapter 5
On-Chip Integration of Eight Directional Droplet
Ejectors to Ink a Spot
This chapter describes an on-chip integration of an acoustic ejector array
consisting of eight directional droplet ejectors that was designed to ink a spot with
eight different droplets without having to move the ejectors. Each of the eight
directional ejectors consistently ejects uniform droplets in diameter of 51 μm with a
directional angle about 17° with respect to the normal direction of the liquid surface.
When a glass substrate was placed 8 mm away from the ejector array, all the ejected
droplets from the 8 ejectors were placed within 399 x 1080 μm
2
area. If we exclude
two ejectors which had bad alignment with others due to the low viscosity glue on
the acoustic lens pattern, all the 6 droplets were placed within 238 x 380 μm
2
area.
5.1 Working Principle
The acoustic droplet ejector is driven by longitudinal acoustic waves produced
from thickness mode resonance of a PZT substrate that are focused through acoustic
Fresnel lens with air-reflector [1] built on the PZT surface. The conventional
acoustic droplet ejectors, which eject droplets in vertical direction, have Fresnel lens
consisting of multiple numbers of annular rings and circular top and bottom
electrodes that covers whole lens area. Because of the symmetric annular Fresnel
74
lens, in-place direction acoustic pressure (or force) cancels out each other, and only
vertical acoustic pressure remains. MATLAB simulation was performed to calculate
vertical and in-plane particle displacement due to the focused acoustic beam
generated from a transducer with circular top and bottom electrodes. As we can see
in Figure 5.1, the in-plane particle displacement at the center of the plot (focal point
of the acoustic Fresnel lens) is zero due to cancelation effect.
One or both of the top and bottom electrodes sandwiching the PZT substrate is
patterned in a pie shape with its sector angle of 90° to generate asymmetric acoustic
pressure at the focal point on the surface of the liquid in the ejector chamber.
MATLAB simulation was also performed to confirm the non-zero in-plane direction
particle displacement generated from the transducer having pie shaped top and
Figure 5.1 MATLAB simulation results showing (a) vertical and (b) in-plane
direction particle displacement due to focused acoustic beam generated from a
transducer with circular top and bottom electrode
75
bottom electrodes with 90° apex angle. At this time, the in-plane direction particle
displacement at the focal point is not cancel out each other because of asymmetric
electrode shape (Figure 5.2). The non-zero in-place direction acoustic pressure is the
reason for a directional droplet ejection (Figure 5.3) [2-3].
Figure 5.2 MATLAB simulation results showing (a) vertical and (b) in-plane
direction particle displacement due to the focused acoustic beam generated from a
transducer with pie-shape top and bottom electrodes with 90° apex angle.
Figure 5.3 Photo of an obliquely ejected droplet after 400 μsec from the ejection
76
5.2 Transducer Design
An ejector array consisting of 8 directional ejectors, each of which operates at
the 3
rd
harmonic resonance of the PZT substrate was designed and fabricated. The
ejector is built on a PZT with two electrodes, on one of which Fresnel air-cavity lens
is formed with parylene D and air pocket inside the parylene. Acoustic waves are
only produced from the PZT region covered with the electrodes on both sides, and
the acoustic waves pass through the lens only where is no air cavity. Asymmetric
acoustic pressure distribution is due to near-field effect of waves after they pass
through the lens [4].
The direction of the droplet ejection with respect to the line perpendicular to
the liquid surface depends mainly on the electrode’s sector angle. Thus, we designed
the eight directional ejectors with varying degrees of the sector angle, in order to ink
a spot with eight different liquid types. However, other ejection condition such as the
curvature of liquid surface, the driving signal frequency, etc. also affect the ejection
direction. It is possible that an ejector ejects droplets at multiple frequencies around
the resonance frequency of the ejector. Therefore, we made sure that the electrodes
and lens shape were symmetric, and each ejector was position at a same relative
position in each ejecting chamber. In addition, the 8 ejectors in an ejector array were
located on circumference of a circle so that distances from the center of the circle to
each ejector are all the same (Figure 5.5). As a result, the ejected droplets from the 8
ejectors inked a point right above of the center point of the circle (Figure 5.4).
77
The reflection of acoustic wave in the ejection chamber is also an important
factor because Fresnel lens focuses acoustic waves through constructive and
destructive wave interference. The reflected acoustic waves also interfere with the
Figure 5.4 Schematic diagram of the array of eight directional ejectors for inking a
spot without moving the ejectors.
Figure 5.5 Photo of 8 Fresnel lenses sitting on circumference of a circle on the
fabricated PZT transducer.
78
acoustic waves coming directly from the transducer and affect the ejection condition.
The silicon ejection chamber shape that affects acoustic reflection condition can
change ejection efficiency and direction. Since two adjacent ejectors have angular
offset of 45°, anisotropic wet etching by KOH on a (100) silicon wafer cannot
provide same acoustic reflection environment to each ejector in the ejection chamber.
Any shape of opening patterned on (100) silicon surface will be squared by KOH
etching with its side lines aligned to (111) direction of the silicon crystalline.
Consequently, we used deep reactive ion etching (DRIE) to provide identical
acoustic wave reflection condition for each ejector in the ejector array (Figure 5.6).
In addition, DRIE allows dense packing of the ejectors that reduces the traveling
distance in the air for the ejected droplets and enhances the accuracy of inking a spot
through reducing ambient-airflow-induced direction change of the ejected droplets.
Figure 5.6 Photos of the fabricated PZT ejector array with silicon microchannels
etched by deep reactive ion etching (DRIE).
79
To have identical ejection condition, not only the ejection chamber shape but
also the relative position of the acoustic lens in its ejection chamber should be same
in each ejector. By aligning the micromachined silicon channel and PZT transducer
in a photo mask aligner, we obtained the relative position of the ejector in its ejection
chamber to be almost same, as shown in Figure 5.7. The PZT transducer and silicon
microchannels were temporarily fixed on the photo mask aligner, and completely
bonded with low viscosity glue epoxy using capillary force [5] outside the photo
mask aligner.
Figure 5.7 Photos of the Fresnel lenses of the ejectors after the silicon wafer with
DRIE-etched microchannels was glued to the lens-containing PZT substrate.
80
5.3 Fabrication and Test Setup
The ejector array was fabricated in following steps as illustrated in Figure 5.8.
To reduce the series resistance from the pad to the active area of the transducer, 0.2
μm-thick additional nickel layer was added onto the pre-deposited 0.1 μm-thick
nickel layer on both sides of a 127 μm-thick PZT sheet (PSI-5A4E, Piezo Systems)
by e-beam metal evaporator. After patterning top and bottom electrodes on the PZT
sheet (Figure 5.8a), photoresist (AZ 5214-E, 30 sec @1200 rpm, 80 mJ/cm
2
) was
patterned as a sacrificial layer, on which 3.5 μm-thick parylene D layer was
deposited as a Fresnel lens structure layer (Figure 5.8b). The sacrificial photoresist
was then released in acetone to form air gaps through release holes with 30 μm
diameter, and additional 7 μm-thick parylene D layer was deposited to seal the
release holes (Figure 5.8d). A microchannel-containing 400 μm-thick silicon wafer
was temporarily glued on the PZT sheet (Figure 5.8e) on a photo mask aligner, and
then low viscosity epoxy resin was filled into the gap between silicon wafer and PZT
sheet by capillary force and cured for 24 hours at 80°C. Finally, a micromachined
silicon wafer was added as the top cover, and the whole stack of wafers was diced
with a dicing saw.
To actuate the fabricated ejector array, pulsed sinusoidal waves were applied to
the ejectors (either one by one, or simultaneously). A pulse generator generated
square pulses with pulse repetition frequency (PRF) of 60 Hz, which triggered RF
signal generator to produce 10 V
peak-to-peak
pulsed sinusoidal signals. A delay control
circuit fed the square pulses to LED with an adjustable fixed delay to illuminate the
81
scene to capture the ejection process only at some desired period of times, as
typically done with optical strobing.
A charge coupled device (CCD) camera (SONY SSC-DC54A) attached at the
end of a Nikon long-working-distance microscope lens captured the scenes, and sent
the video signal to a computer to record the signal. A slide glass with 2 mm grid
mark was placed above the ejector array to collect the ejected droplets. A still camera
with a macro lens was placed right above slide glass to record the droplet location,
while the CCD camera took the video from the side of the ejector array to observe
the ejection in real-time (Figure 5.9).
Figure 5.8 Brief fabrication steps of the ejector array
82
Figure 5.9 Measurement setup of ejector array’s inking a spot operation
Figure 5.10 Photos of (a) top-view and (b) bottom-view of the fabricated ejector
array
83
5.4 Measurement Results and Discussions
The performance of the Ejector 1 and Ejector 2 was tested using DI water
(Figure 5.11). Two droplets broke off simultaneously from their own bulk liquids 50
μsec after the rising edge of a pulse (Figure 5.11a). They then approach each other.
At around 2,000 μsec, they came close to each other in air (Figure 5.11d). The
droplet size (51 μm) was much smaller than the distance between the two ejectors
(4.2 mm), yet the multiple droplets ejected by the eight adjacent ejectors still came
close to each other in a three-dimensional space owing to the similar
ejection/traveling speed and good directionality.
Figure 5.11 Photos of droplet ejection process captured at (a) 50 μsec, (b) 300 μsec,
(c) 900 μsec, and (d) 2,000 μsec after actuating Ejector 1 and Ejector 2.
84
Since each ejector in the array can eject droplets at various frequencies near the
resonant frequency, the ejectors were tested to find out the frequency at which the
ejector makes stable droplet ejection in the desired direction. Typically, a minimum
of 7 ~ 18 μsec (rather wide range) pulse width was required to eject droplets from the
eight different ejectors (Table 5.1). The wide variation of optimal pulse width is due
to relatively poor fabrication precision and PZT thickness uniformity.
With a glass substrate placed 8 mm away from the ejector array, each of the
eight ejectors was actuated one by one to eject 180 ~ 240 water droplets forming
water bump on the glass substrate, which has 2 mm x 2 mm white grid on it. From
the photos of the 8 water bumps (Figure 5.12), the center point of each water bump is
plotted having the lower left corner of grid square as the origin of the plot in Figure
5.13. The ejected droplets from the 6 ejectors fall into a very small area of 238 x 380
μm
2
[6], which can certainly be made much smaller with better alignment accuracy
during the fabrication and/or a PZT substrate with more uniform thickness.
Table 5.1 Ejection condition for the eight directional ejectors of the ejector array
Ejector Frequency (MHz) Pulse Width (μsec)
Ejector 1 50.940 10
Ejector 2 51.346 7.3
Ejector 3 51.813 10
Ejector 4 52.447 18
Ejector 5 51.723 11
Ejector 6 51.127 9
Ejector 7 50.940 10
Ejector 8 51.213 10
85
Figure 5.12 Photo of the water bump collected on a glass slide after the 8 ejectors of
the ejector array ejected about 200 droplets onto the glass slide. Red dots indicate the
center of the water bumps.
Figure 5.13 Plot of the center locations of the water bumps (on the glass slide)
formed by the droplets ejected by the 8 ejectors without moving the ejectors, showing
close proximity of the droplet placements.
86
5.5 Summary
Non-vertical ejection is one of the grate advantages of nozzleless ejectors. An
ejector array consisting of nozzleless droplet ejectors is capable of inking a spot with
various kinds of liquid without ejector movement. The ejector array especially
enhances the system throughput when it is utilized in the in-situ peptide synthesis
system using Spot
technique described in Chapter 4.
An ejector array consisting of 8 directional acoustic ejectors operating in their
third harmonic frequency of the PZT transducer was fabricated and tested.
Considering the small droplet size of 51 μm and relatively long distance of 8 mm
from each ejector of the ejector array to the inking spot on the glass slide, the ejector
array showed a very good accuracy aiming a spot. Droplets ejected from the 6
ejectors out of the 8 ejectors in the ejector array were placed within 238 x 380 μm
2
area. With the ability to eject droplets of 8 different kinds of liquids in eight different
directions with great accuracy, the ejector array enables inking of a spot with
multiple droplets without any mechanical movement and alignment.
87
Chapter 5 References
[1] C.Y. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing
Air-Reflectors,” the 19
th
IEEE International Conference on Micro Electro
Mechanical Systems, pp. 170-173, 2006.
[2] J.W. Kwon, H, Yu, Q. Zhou, and E.S. Kim, “Directional Ejection of Liquid
Droplets by Sectored Self-Focusing Acoustic Transducers Built on ZnO and
PZT,” Journal of Micromechanics and Microengineering, vol. 16, pp. 2697-
2704, 2006.
[3] J.W. Kwon, Q. Zhou, and E.S. Kim, “Directional Ejection of Liquid Droplets
through Sectoring Half-Wave-Band Sources of Self-Focusing Acoustic
Transducer,” IEEE International Micro Electro Mechanical Systems
Conference, Las Vegas, Nevada, January 20-24, pp. 121-124, 2002.
[4] C.Y. Lee, H. Yu, and E.S. Kim, “Nanoliter droplet coalescence in air by
directional acoustic ejection,” Applied Physics Letter, vol. 89, pp. 223902-
223902-3, 2006.
[5] J.W. Kwon, H. Yu, and E.S. Kim, “Film transfer and bonding technique for
covering single-chip ejector array with microchannels and reservoirs,” Journal
of Microelectromechanical Systems, vol. 14, pp. 1399-1408, 2005.
[6] Y. Choe, L. Wang, and E.S. Kim, “On-chip Integration of Eight Directional
Droplet Ejectors for Inking a Spot without Ejector Movement,” Transducers ’11,
IEEE International Solid-State Sensors and Actuators, Beijing, China, June 5-9,
pp. 2948-2951, 2011.
88
Chapter 6
Acoustic Tweezers with Multi-foci Fresnel Lenses
Contactless particle trapping can be realized using both optical waves and
acoustic waves. Unlike optical tweezers, acoustic tweezers is less dependent on the
wave transfer media (optical tweezers need transparent media), has less energy
density, and therefore, is safer to be applied to living cells than optical tweezers.
This chapter describes an acoustic tweezers consisting of a multi-foci Fresnel
lens on a 127 μm thick PZT sheet to demonstrate microparticle trapping ability on a
liquid surface by a single ultrasonic transducer. The multi-foci Fresnel lens was
designed to have similar working mechanism as that of an axicon lens to generate an
acoustic Bessel beam, which has negative axial radiation force capable of trapping
one or more microparticles. The fabricated acoustic tweezers successfully trapped
lipid particles ranging in diameter from 50 to 200 μm and microspheres ranging in
diameter from 70 to 90 μm and from 500 to 600 μm at a distance of 2 to 5 mm from
the tweezers without any contact between the transducer and microparticles.
6.1 Negative Axial Radiation Force in a Bessel Beam
A Bessel beam is well known for its non-diffractive and self-healing
characteristics [1]. The self-healing means that the beam can be partially obstructed
at one point, but will re-form at a point further down the beam propagation axis. A
Bessel beam also has negative axial radiation force under certain conditions. Above
89
three characteristics make a Bessel beam very attractive for optical and acoustic
tweezers applications.
When a Fresnel lens is designed to have multiple focal points, the acoustic
waves passing through the lens produces a Bessel beam with a micron scale region
where the radiation force is negative, and microparticles can be trapped. The wave
equation ψ
B
for a scalar-wave Bessel beam is an asymmetric solution of the free-
space wave equation.
( )
( )
( √
) (6-1)
where, ψ
0
, κ, J
0
, and μ are the wave amplitude, axial wave number, zeroth-
order Bessel function, and radial wave number, respectively.
Marston [2] mathematically proved that negative axial radiation force exists in
a Bessel beam under certain conditions as shown experimentally by Whitworth [3].
To explain the conditions for the negative axial radiation force, Marston defined
parameter domain (k, α, β), where k and α are the wave number and radius of
trapping particle, respectively, while β is a cone angle. The cone angle β is an
important parameter in the characterization of a Bessel beam, and linked to the
parameters of Eq. 6-1 by
( )
( ) (6-2)
also,
(
)
(6-3)
where, ω and c0 are angular frequency and the phase velocity of the acoustic
wave in the liquid, respectively.
90
6.2 Lens Designs
6.2.1 Aluminum Axicon Lens
The conventional way to generate a Bessel beam is to use axicon lens. An
acoustic axicon lens was designed with T2024 aluminum alloy to compare the
performance with multi-foci Fresnel lens. To have the cone angle β=60°, the axicon
lens angle α was set to 63° (Figure 6.1). To form an acoustic tweezes, 127 μm thick
PZT sheet with 0.4 μm top and bottom nickel electrodes were combined with
aluminum axicon lens using low viscosity glue to minimize the glue thickness.
6.2.2 Multi-foci Fresnel Lens
Observing the working mechanism of an axicon lens, we can see that the wave
closer to the lens center is focused at a shorter distance from the lens. In the other
ends, the wave farther from the lens center is focused at longer distance from the lens
Figure 6.1 Schematics of the aluminum alloy axicon lens: (a) top view (b) side view
91
as shown in Figure 6.2. A Bessel beam is formed in the region where the focused
wave is uniformly distributed.
To simulate the behavior of an axicon lens described above, a multi-foci
Fresnel lens with an air-reflector [4] was designed. The Fresnel lens has 7 rings in
total. The 2 inner most rings, the next 2 rings, and the remaining 3 rings have focal
lengths of 830 μm, 860 μm, and 890 μm, respectively (Figure 6.3) to have linear
distribution of focused acoustic beam along the acoustic wave propagation axis. As a
control experiment, an acoustic tweezers with Fresnel lens having only one focal
length of 800 μm and same number of rings was fabricated. When the transducer
with single focal length of 800 μm was actuated in DI water, particles freely floating
above the transducer were immediately expelled from the region due to positive axial
radiation force and the stirring effect. Therefore, no particle was trapped by the
single focal point transducer.
Figure 6.2 Ray trajectories of the acoustic waves going through an axicon lens
92
6.3 Device Fabrication
6.3.1 Acoustic Tweezers with Aluminum Axicon Lens
The fabrication steps of an acoustic tweezers with aluminum axicon lens are
very simple. Firstly, 0.4 μm thick top and bottom electrodes of a 127 μm thick PZT
transducer were patterned to form acoustic transducers. A 400 μm thick bulk-
micromachined silicon wafer was attached on back side of the transducer to
mechanically support. Aluminum alloy axicon lens was fabricated in USC machine
shop using milling machine to have the 63° cone angle. The aluminum axicon lens
was attached on front side of the transducer using low viscosity glue, and electric
pads were soldered to external wires. Since acoustic tweezers operated in liquid, 10
μm thick parylene D layer was deposited to encapsulate whole tweezers for electric
insulation between top and bottom electrodes.
Figure 6.3 Side-view schematic of the multi-foci Fresnel lens
93
Figure 6.4 Fabrication steps of the acoustic tweezers with an aluminum alloy axicon
lens.
Figure 6.5 Photos of (a) aluminum alloy axicon lens (b) acoustic tweezers with the
fabricated aluminum alloy axicon lens
94
6.3.2 Acoustic Tweezers with Multi-foci Fresnel Lens
Acoustic tweezers with multi-foci Fresnel lens was fabricated as follow. To
reduce the series resistance of electrode, a 0.3 μm thick additional nickel layer was
deposited using e-beam metal evaporator onto the 0.1 μm thick pre-deposited nickel
layer on both sides of a 127 μm thick PZT sheet. Top and bottom nickel layer was
patterned with nickel etchant TFG (Transene, Inc.) to form transducer. AZ5214
photoresist was spun on spinner with 1.2 kRPM rotational speed and patterned to
form about 2 μm thick multi-foci Fresnel lens structure as a sacrificial layer. 3.5 μm
thick parylene D layer was deposited as a lens structural layer, and the sacrificial
photoresist was released through release holes with 30 μm diameter in acetone to
form air gaps. Additional 7 μm thick parylene D layer was deposited to seal the
release holes. Finally, bulk-micromachined silicon was glued to the transducer as a
structural support layer and as an acoustic beam window.
Figure 6.6 (a) SEM photo of the fabricated multi-foci Fresnel lens on a PZT sheet
(b) Photo of a packaged acoustic tweezers with the multi-foci Fresnel lens
95
6.4 Particle Trapping Measurement Setup and Results
6.4.1 Measurement Setup
For a low energy in trapping microparticles, a pulsed 17.9 MHz sinusoidal
signal, instead of continuous sinusoidal signal, was applied to the fabricated acoustic
tweezers with 10 – 20 kHz pulse repetition frequency (PRF). The pulse width was
fixed to 2 μsec, and the sinusoidal signal was amplified to have 160 V
peak-to-peak
voltage amplitude. Both acoustic tweezers with aluminum alloy axicon lens and with
multi-foci Fresnel lens were submerged in DI water and fixed to a manual XYZ
Figure 6.7 Brief fabrication steps of the acoustic tweezers built on a 127 μm thick
PZT sheet with multi-foci Fresnel lens.
96
moving stage. The movement of the microparticles was observed through a CCD
attached a long range microscope, and the images and videos captured by the CCD
were recorded with a computer.
Lipid particles and polystyrene particles were used for tapping particles. To
harvest lipid particles, an acoustic ejector ejected olive oil droplets into the air, and
the ejected droplets were collected with beaker filled with DI water. Since the
droplet size is constant as 70 μm in diameter, most of the collected lipids had
diameters of 70 μm. However, some lipids in collecting beaker merged to each other
and form larger lipids having diameter up to 200 μm. Polystyrene microspheres had
diameters in range of 75 ~ 90 μm with density of 0.99 g/cm
3
. Both lipid particles and
polystyrene particles were freely floating on water surface when they were in test.
Figure 6.8 Schematic of the setup to observe the trapping of microparticles by the
fabricated acoustic tweezers
97
6.4.2 Lipid Particle Trapping
The fabricated acoustic tweezers, both tweezers based on the multi-foci Fresnel
lens and with the aluminum alloy axicon lens, were tested to trap lipid particles
ranging from 70 ~ 200 μm in diameter and microspheres ranging from 75 ~ 90 μm in
diameter [5,6]. As the actuated tweezers produced acoustic waves and stirred the
water as well as the microparticles in/on the water, the microparticles circled around
the tweezers. Once a lipid particle hit the location where the Bessel beam was
generated, the lipid particle firmly trapped to the spot and held there even when
another lipid particle hit the trapped lipid particle (Figure 6.9). And we were able to
move the trapped particle by moving the transducer over a wide range (limited only
by the movable range of the transducer in the test apparatus).
Figure 6.9 Photo of manual XYZ control
98
Figure 6.10 Photos taken at different times showing that a trapped lipid particle
holds its position even when hit by another moving lipid particle.
Figure 6.11 A large lipid particle with 200 μm diameter was trapped by the
fabricated acoustic tweezers.
99
6.4.3 Polystyrene Microparticle Trapping
The acoustic tweezers with top silicon cover was also test to capture
polystyrene microparticle with diameter from 75 ~ 90 μm. The silicon top cover
works as an acoustic beam window. Due to the constructive and destructive
interfering mechanism of Fresnel lenses, beam profile without acoustic beam
window can be wide. To have narrow capturing point, a bulk-micromachined silicon
acoustic beam window with 600 μm x 600 μm opening that is aligned to the center of
Fresnel lens was placed on top of the acoustic tweezers. The acoustic tweezers with
silicon acoustic beam windows accurately capture only one or two polystyrene
microparticles at its capture point (Figure 6.12)
An acoustic tweezers without silicon acoustic beam window was also
fabricated. Instead of having silicon structural layer, 2 mm thick acrylic plate
mechanically supports the acoustic tweezers from the back side. The PZT transducer
of the acoustic tweezers without acoustic beam windows was designed and
fabricated in the same manner with that of acoustic tweezers with the acoustic beam
window. When the acoustic tweezers without the window was actuated, a group of
polystyrene microparticles with diameter of 75 ~ 90 μm was attracted right above the
center of multi-foci Fresnel lens and kept trapped even when the acoustic tweezers
was moved (Figure 6.13). Though same size and same types of microparticles were
used testing both acoustic tweezers with and without acoustic beam window, the
acoustic tweezers trapped microparticles in quite different way due to the beam
profile difference.
100
Figure 6.12 Photo of trapped microparticle with 80 μm diameter by an acoustic
tweezers with silicon window
Figure 6.13 A group of polystyrene microspheres with 80 μm diameters was
trapped by an acoustic tweezers without silicon window
101
6.4.4 Beam Profile Measurement
The beam profile of acoustic tweezers without silicon beam shaping window
was measured with a hydrophone (PAHPM04/1) as shown in Figure 6.14. The device
was actuated with pulsed sinusoidal signal with 200 Hz pulse repetition frequency
(PRF), and each pulse had 10 cycles of 17.6 MHz sinusoidal waves. The output
voltage of the function generator (Tektronix, AFG3252) was set to 2 V
peak-to-peak
, and
the gain of the power amplifier (AR 75A250) was set to minimum to have voltage
amplitude of 37V
peak-to-peak
. To measure pressure distribution along the z-axis, the
hydrophone was aligned to the center of the multi-foci Fresnel lens and scanned
along z-axis from 2 mm to 6.3 mm with 20 μm position scanning interval. According
to the axial scan results of the hydrophone, several local peaks were observed at 2.9
mm, 4.0 mm, and 5.8 mm away from the Fresnel lens.
Figure 6.14 Photo of hydrophone measurement setup (a) projection view (b) side
view
102
Lateral scans were also performed at 2.9 mm, 4.0 mm, and 5.8 mm away from
the transducer surface to see lateral beam profile. The hydrophone was moved about
1.2 mm along the center line of the multi-foci Fresnel lens with 5μm interval. Since
the acoustic pressure depends only on the distance from the center of the lens to
measuring point, any scanning line should give same result if the scanning passes the
center point of the lens.
Since the trapping occurs right above the center of the lens pattern, the beam
profile at the center is most important. When measured 2.9 mm and 4.0 mm away
from the transducer, the center beam profile had multiple peaks (Figure 6.15 and
Figure 6.16) while the beam profile measured at 5.8 mm away showed single peak
(Figure 6.17).
Figure 6.15 Hydrophone measurement results along z-axis (wave propagation
direction)
103
Figure 6.16 Hydrophone measurement results along lateral direction at 2.9 mm away
from the acoustic tweezers
Figure 6.17 Hydrophone measurement results along lateral direction at 4.0 mm away
from the acoustic tweezers
Figure 6.18 Hydrophone measurement results along lateral direction at 5.8 mm away
from the acoustic tweezers
104
6.4.5 Zebra Fish Egg Trapping
Since it turned out that the acoustic tweezers without acoustic beam window
has wide beam width, polystyrene microparticles with diameter of 500 ~ 600 μm was
tested for trapping. The acoustic tweezers was submerged in DI water 3 mm below
the water surface, and the microparticles were freely floating on the water (the
density of microparticles is 0.99 g/cm
3
). When the acoustic tweezers was actuated,
two microparticles were firmly trapped and followed the movement of acoustic
tweezers (Figure 6.19).
Zebra fish eggs are known to have egg diameters of about 500 μm. A fresh
zebra fish which is only few hours old (i.e. the egg was delivered few hours before
the experiment) was placed on the acoustic tweezers to see if the tweezers can trap
the zebra fish egg. Since the zebra fish egg is heavier than DI water and the acoustic
tweezers can only trap particles above the lens, the zebra fish egg was placed right
on the center of the multi-foci Fresnel lens of the acoustic tweezers.
When the acoustic tweezers was actuated with pulsed sinusoidal signal with 10
kHz PRF, 2 μsec of pulse width, and 18 MHz driving signal, the zebra fish egg was
pushed upward until it floats on the water surface and trapped by the acoustic
tweezers holing its position (Figure 6.20). When the acoustic tweezers was moved,
the trapped zebra fish egg was followed the tweezers movement. Since the zebra fish
eggs are heavier than lipids and polystyrene microparticles, the holding of zebra fish
egg was weaker than that of other microparticles.
105
Figure 6.19 Two polystyrene microparticles with diameter of 500 μm are trapped by
the acoustic tweezers.
Figure 6.20 A Zebra fish egg was successfully trapped by the acoustic tweezers
with multi-foci Fresnel lens.
106
6.5 Summary
An ultrasound Bessel beam was generated using acoustic transducers with
multi-foci Fresnel lens. The multi-foci Fresnel lens successfully simulated the
behavior of conventional axicon lens and produced negative axial pressure to capture
microparticles. The formation of ultrasound Bessel beam was confirmed by
hydrophone measurements and also by actual microparticle trapping experiments.
The ultrasound acoustic tweezers with multi-foci Fresnel lens showed excellent
performance in trapping lipids, polystyrene microparticles, and even a Zebra fish egg.
For performance comparison, aluminum axicon lens was fabricated and tested
for microparticle trapping. As expected, the acoustic tweezers with aluminum alloy
axicon lens showed poor performance due to the difficulty of miniaturizing
geometrical lens shape. The acoustic tweezers was indeed able to trap polystyrene
microparticle with diameter of 70 μm. However, the acoustic tweezers shortly lost
control on the microparticle, and it was very hard to repeat the result.
Though Bessel beam has great characteristics applicable to acoustic tweezers,
acoustic tweezers hardly used the Bessel beam due to the difficulty of miniaturizing
axicon lenses with good quality. A virtual acoustic axicon lens employing multi-foci
Fresnel lens was designed, fabricated for the first time, and successfully
demonstrated the formation of an acoustic Bessel beam by trapping various kinds of
microparticles.
107
Chapter 6 References
[1] F.O. Fahrbach, P. Simon, and A.Rohrbach, “Microscopy with self-
reconstructing beams,” Nature Photonics, November, vol. 4, pp. 780-785, 2010.
[2] P.L. Marston, “Axial radiation force of a Bessel beam on a sphere and direction
reversal of the force,” Journal of Acoustic Society of America, vol. 120, pp.
3518-3524, 2006.
[3] G. Whitworth, “Particle column formation in a stationary ultrasonic field,”
Journal of Acoustic Society of America, vol. 91, pp. 79-85, 1992.
[4] C.Y. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing
Air-Reflectors,” the 19
th
IEEE International Conference on Micro Electro
Mechanical Systems, pp. 170-173, 2006.
[5] Y. Choe, J.W. Kim, K.K. Shung, and E.S.Kim, “Ultrasound Microparticle
Trapping by Multi-Foci Fresnel Lens,” Joint Conference of the IEEE
International Frequency Control Symposium and European Frequency and
Time Forum, San Francisco, CA, May 1-5, 2011.
[6] Y. Choe, J.W. Kim, K.K. Shung, and E.S. Kim, “Microparticle Trapping in An
Ultrasonic Bessel Beam,” Applied Physics Letter, vol. 99, pp. 233704, 2011.
108
Chapter 7
Conclusion and Future Directions
Piezoelectric MEMS based on PZT sheet have been designed, fabricated, and
tested for various airborne and liquid applications. Microspeakers based on PZT
bimorph were demonstrate to outperform a commercial cell-phone electromagnetic
speaker in terms of the sound pressure output for a given power consumption.
Microfluidic pumps based on acoustic streaming were fabricated to circulate liquid
with very high volume transfer rate, while acoustic droplet ejectors were improved to
ink a spot with multiple droplets with precise control over the ejection direction and
dispensing volume. Also, ultrasound waves were focused to form a Bessel beam to
firmly trap various kinds of microparticles and zebra fish eggs. These results clearly
indicate that integration of PZT sheet and micromachined silicon structures leads to
many innovative acoustic MEMS for some impactful applications. In spite of these
impressive results, there is room for improvement in the devices, and the followings
describe some of potential future directions.
7.1 Thinner Microspeaker on Round PZT Diaphragm
Integration of voltage boosting circuit is not difficult with cutting-edge IC
technology up to a certain level. However, 190V
peak-to-peak
actuation voltage may not
be in the comfort zone for mobile devices. Mechanical polishing has already been
109
applied to the microspeaker to have smoother PZT surface, but we can more
aggressively apply the mechanical polishing technique to have thinner PZT bimorph
diaphragm. The actuation voltage will be scaled with linear relationship to the
reduced diaphragm thickness.
Though the top and bottom electrode were patterned with great care, some
areas near the four corners of the square diaphragm do not contribute to the sound
generation because of the zero displacement in those areas. This is inherent
characteristics of four edge clamped diaphragm, and it lowers area utilizing
efficiency. A Bimorph diaphragm having round boundary condition, which can be
obtained by using dry silicon etching technique, can solve this issue. Except the zero
displacement area between center and edge of the diaphragm due to the opposite
direction of stress in those areas, all other area has bending moment and contributes
to sound generation due to the symmetric geometry of the diaphragm.
7.2 Micropumps with Non-vertical Reflector
In our past micropump experiment, vertical acrylic reflector was formed by
laser cutter to reduce undesirable interference between adjacent lenses. However, the
selectivity between passing desirable direction waves and reflecting undesirable
direction waves may sit in very narrow margin with vertical reflector. Having non-
vertical reflector which is tilted toward liquid flow direction will certainly reduce the
possibility for the reflector to interfere the acoustic waves proceeding in desirable
direction, and increase the reflection efficiency to the acoustic waves proceeding in
110
undesirable direction. Double side KOH etching of a (100) silicon wafer can produce
such slanted reflector, and 3-D printer is another option to realize the reflector with
much wider variety of tilting angle.
7.3 Protein Synthesis with Ejector Array
The vertical acoustic ejector needs to be replaced with an array of ejectors to
eject different kinds of amino acid solution onto an active spot on a modified glass
substrate, because the process time of synthesizing peptides consisting of various
kinds of amino acids is long due to the ejector replacement and cleaning. Replacing
the single vertical ejector with an ejector array consisting of 8 or even larger number
of ejectors will significantly reduce the process time and the system throughput.
Liquid transfer channel routing would be a new issue to be addressed in case of using
ejector arrays. Possible candidates for resolving the liquid channel routing problem
are tubing technique and multi-layer liquid channel.
7.4 Higher Harmonic Acoustic Tweezers
Acoustic tweezers operating at 18 MHz fundamental resonance frequency of
PZT substrate successfully trapped microparticles having wide range of diameters.
However, the tweezers trapped a group of microparticles when operated without a
silicon window. If we try to capture microparticles which have much smaller size
than acoustic wavelength, it is highly possible for the acoustic tweezers working in
its fundamental resonance frequency to trap a group of small microparticles rather
111
than trap a single particle even with a silicon window. The acoustic tweezers
demonstrated in chapter 6 generates acoustic waves having wave length of 80 μm
which is much larger than normal animal cells. To expand the operation range of
single particle trapping acoustic tweezers further, it is worth to try acoustic tweezers
working in higher harmonic frequency to precisely trap microparticles which have
about 10 μm or smaller diameter.
112
Bibliography
A.
F.L. Ayatollahi and B.Y. Majlis, “Materials Designed and Analysis of Low-Power
MEMS Microspeaker Using Magnetic Actuation Technology,” Advanced Materials
Research, vol. 74, pp. 243-246, 2009.
B.
V. Benetics, A. Shooshtari, and P. Foroughi, “A source-integrated micropump for
cooling of high heat flux electronics,” Nineteenth Annual IEEE Semiconductor
Thermal Measurement and Management Symposium, San Jose, CA, March 11-13, pp.
236-241, 2003.
H.E. Blackwell, “Hitting the SPOT: small-molecule macroarrays advance
combinatorial synthesis,” Current Opinion in Chemical Biology, vol. 10, pp. 203-212,
2006.
M.D. Bowman, R.C. Jeske, and H.E. Blackwell, “Microwave-Accelerated SPOT-
Synthesis on Cellulose Supports,” Organic Letters, vol. 6, pp. 2019-2022, 2004.
C.
P.C.P. Chao, C.W. Chiu, and Y. Hsu-Pang, “Magneto-Electrodynamical Modeling
and Design of a Microspeaker Used for Mobile Phones With Considerations of
Diaphragm Corrugation and Air Closures,” IEEE Transactions on Magnetics, vol. 43,
pp. 2585-2587, 2007.
Y.C. Chen, W.T. Liu, T.Y. Chao, and Y.T. Cheng, “AN OPTIMIZED CU-NI
NANOCOMPOSITE COIL FOR LOW-POWER ELECTROMAGNETIC
MICROSPEAKER FABRICATION,” Transducers ’09, IEEE International
Conference on Solid-State Sensors and Actuators, Denver, CO, June 21-25, pp. 25-
28, 2009.
Y.C. Chen and Y.T. Cheng, “A LOW-POWER MILLIWATT
ELECTROMAGNETIC MICROSPEAKER USING A PDMS MEMBRANE FOR
HEARING AIDS APPLICATION,” IEEE International Micro Electro Mechanical
Systems Conference, Cancun, Mexico, January 23-27, pp. 1213-1216, 2011.
Y. Choe, S.J. Chen, and E.S. Kim, “Bimorph Diaphragm Formed by Two PZT
Sheets on Micromachined Silicon for Sound Generation,” 6
th
International
Conference and Exhibition on Device Packaging, Scottdale/Fountain Hills, AZ,
March 9-11, pp. 269-270, 2010.
113
Y. Choe, S.J. Chen, and E.S. Kim, “High Fidelity Loud Microspeaker Based on PZT
Bimorph Diaphragm,” Technical Proceeding of the 2010 NSTI Nanotechnology
Conference & Expo, Anaheim, CA, June 21-25, vol. 2, pp. 316-319, 2010.
Y. Choe, L. Wang, and E.S. Kim, “On-chip Integration of Eight Directional Droplet
Ejectors for Inking a Spot without Ejector Movement,” Transducers ’11, IEEE
International Solid-State Sensors and Actuators, Beijing, China, June 5-9, pp. 2948-
2951, 2011.
Y. Choe, J.W. Kim, K.K. Shung, and E.S.Kim, “Ultrasound Microparticle Trapping
by Multi-Foci Fresnel Lens,” Joint Conference of the IEEE International Frequency
Control Symposium and European Frequency and Time Forum, San Francisco, CA,
May 1-5, 2011.
Y. Choe, J.W. Kim, K.K. Shung, and E.S. Kim, “Microparticle Trapping in An
Ultrasonic Bessel Beam,” Applied Physics Letter, vol. 99, pp. 233704, 2011.
E.
S.A. Elrod, B. Hadimioglu, B.T. Khuri-Yakub, E.G. Rawson, E. Richley, N.N.
Mansour, and T.S. Lundgren, “Nozzleless droplet formation with focused acoustic
beams,” Journal of Applied Physics, vol. 69, pp. 3341-3347, 1989.
X. Espanel, R.H. Huijsduijnen, “Applying the SPOT peptide synthesis procedure to
the study of protein tyrosine phosphatase substrate specificity: probing for the
heavenly match in vitro,” Methods, vol. 35, pp. 64-72, 2005.
F.
F.O. Fahrbach, P. Simon, and A.Rohrbach, “Microscopy with self-reconstructing
beams,” Nature Photonics, vol. 4, pp. 780-785, 2010.
R. Frank, “Spot-Synthesis: An Easy Technique for the Positionally Addressable,
Parallel Chemical Synthesis on a Membrane Support,” Tetrahedron, vol. 48, pp.
9217-9232, 1992.
R. Frank, “The SPOT-synthesis technique. Synthetic peptide arrays on membrane
supports—principles and applications,” Journal of Immunological Methods, vol. 267,
pp. 13-26, 2002.
G.
D.G. Grier, “A revolution in optical manipulation,” Nature, vol. 429, pp. 810-816,
2003.
114
S. Guo and J. Wang, “A Novel Type of Micropump Using Solenoid Actuator for
Biomedical Applications,” 2007 IEEE International Conference on Robotics and
Automation, Roma, Italy, April 10-14, pp. 654-659, 2007.
S. Guo, J. Wang, and Q. Pan, “Solenoid Actuator-based Novel Type of Micropump”
IEEE International Conference on Robotics and Biomimetics, Kunming, China,
December 17-20, pp.1281-1286, 2006.
S. Guo and J. Oohira, “A novel type of micropump using SMA actuator for
microflow application,” Proceedings of 2003 International Symposium on
Micromechatronics and Human Science, Nagoya, Japan, October 19-22, pp. 45-50,
2003.
H.
N. Heine, T. Ast, J. Schneider-Mergener, U. Reineke, L. Germeroth, and H.
Wenschuh, “Synthesis and screening of peptoid arrays on cellulose membranes,”
Tetrahedron, vol. 59, pp. 9919-9930, 2003.
K. Hilpert, D.FH. Winkler, R.EW. Hancock, “Peptide arrays on cellulose support:
SPOT synthesis, a time and cost efficient method for synthesis of large numbers of
peptides in a parallel and addressable fashion” Nature Protocols, vol. 2, pp. 1333-
1349, 2007.
D. Huang and E.S. Kim, “Micromachined Acoustic-Wave Liquid Ejector”,
IEEE/ASME Journal of Microelectromechanical Systems, vol. 10, pp. 442-449, 2010.
G.Y. Hwang, H.G. Kim, S.M. Hwang, and B.S. Kang, “Analysis of Harmonic
Distortion Due to Uneven Magnetic Field in a Microspeaker Used for Mobile
Phones,” IEEE Transactions on Magnetics, vol. 38, pp. 2376-2378, 2002.
S.M. Hwang, G.Y. Hwang, J.H. Kwon, H.J. Lee, and B.S. Kang, “Performance
Comparison Between Circular and Elliptical Type Microspeakers for Cellular
Phones,” IEEE Transactions on Magnetics, vol. 39, pp. 3256-3258, 2003.
S.M. Hwang, J.H. Kwon, and K.S. Hong, “Development of Woofer Microspeaker
Used for Cellular Phones,” IEEE Transactions on Magnetics, vol. 41, pp. 3808-3810,
2005.
S.M. Hwang, H.J. Lee, K.S. Hong, B.S. Kang, and G.Y. Hwang, “New Development
of Combined Permanent-Magnet Type Microspeakers Used for Cellular Phones,”
IEEE Transactions on Magnetics, vol. 41, pp. 2000-2003, 2005.
115
K.
H. Kim, A.A. Astle, K. Najafi, L.P. Bernal, P.D. Washabaugh, and F. Cheng, “Bi-
directional Electrostatic Microspeaker with Two Large-Deflection Flexible
Membranes Actuated by Single/Dual Electrodes,” Sensors 2005, IEEE Conference
on Sensors, Irvine, CA, October 31-November 3, pp. 89-92, 2005.
H.J. Kim, K. Koo, S.Q. Lee, K.H. Park, and J. Kim, “High Performance
Piezoelectric Microspeakers and Thin Speakers Array System,” ETRI Journal, vol.
31, pp. 680-687, 2009.
D.H. Kim, D.S. Shin, and Y.S. Lee, “Spot arrays on modified glass surface for
efficient SPOT synthesis and on-chip bioassay of peptides,” Journal of Peptide
Science, vol. 13, pp. 625-633, 2007.
E.S. Kim, “Integrated microphone with CMOS circuits on a single chip”, Ph.D
dissertation, University of California, Berkeley, 1990.
S.C. Ko, Y.C. Kim, S.S. Lee, S.H. Choi, and S.R. Kim, “Micromachined
piezoelectric membrane acoustic device,” Sensors and Actuators A: Physical, vol.
103, pp. 130-134, 2003.
M. Koch, A.G.R. Evans, and A. Brunnschweiler, “The dynamic micropump driven
with a screen printed PZT actuator,” Journal of Micromechanics and
Microengineering, vol. 8, pp. 119-122, 1999.
A. Kramer, U. Reineke, L. Dong, B. Hoffmann, U. Hoffmȕller, D. Winkler, R.
Volkmer-Engert, and J. Schneider-Mergener, “Spot synthesis: observations and
optimizations,” Journal of Peptide Research, vol. 54, pp. 1-9, 1999.
G.H. Kwon, G.S. Jeong, J.Y. Park, J.H. Moon, and S.H. Lee, “A low-energy-
consumption electroactive valveless micropump for long-term biomedical
applications,” Lab on a Chip, vol. 11, pp. 2910-2915, 2011.
J.W. Kwon, S. Kamal-Bahl and E.S. Kim, “Film Transfer and Bonding Technique to
Cover Lab on a Chip,” Transducers ’05, IEEE International Conference on Solid-
State Sensors and Actuators, (Seoul, Korea), June 5-9, pp. 940-943, 2005.
J.W. Kwon, S. Kamal-Bahl, and E.S. Kim, “In-situ DNA Synthesis on Glass
Substrate for Microarray Fabrication Using Self-Focusing Acoustic Transducer,”
IEEE Transactions on Automation Science and Engineering, vol. 3, pp. 152-158,
2006.
116
J.W. Kwon, H, Yu, Q. Zhou, and E.S. Kim, “Directional Ejection of Liquid Droplets
by Sectored Self-Focusing Acoustic Transducers Built on ZnO and PZT,” Journal of
Micromechanics and Microengineering, vol. 16, pp. 2697-2704, 2006
J.W. Kwon, Q. Zhou, and E.S. Kim, “Directional Ejection of Liquid Droplets
through Sectoring Half-Wave-Band Sources of Self-Focusing Acoustic Transducer,”
IEEE International Micro Electro Mechanical Systems Conference, Las Vegas,
Nevada, January 20-24, pp. 121-124, 2002
J.W. Kwon, H. Yu, and E.S. Kim, “Film transfer and bonding technique for covering
single-chip ejector array with microchannels and reservoirs,” Journal of
Microelectromechanical Systems, vol. 14, pp. 1399-1408, 2005
L.
B. Landenberger, H. Höfemann, S. Wadle, and A. Rohrbach, “Microfluidic sorting
of arbitrary cells with dynamic optical tweezers,” Lab on a Chip, vol. 12, pp. 3177-
3183, 2012.
C.M. Lee, J.H. Kwon, K.S. Kim, J.H. Park, and S.M. Hwang, “Design and Analysis
of Microspeakers to Improve Sound Characteristics in a Low Frequency Range,”
IEEE Transactions on Magnetics, vol. 46, pp. 2048-2051, 2010.
C.K. Lee, A.J. Robinson, and C.Y. Ching, “Development of EHD Ion-Drag
Micropump for Microscale Electronics Cooling Systems,” 13
th
International
Workshop on Thermal Investigation of ICs and Systems, Budapest, Hungary,
September 17-19, pp. 48-53, 2008.
C.Y. Lee, H. Yu, and E.S. Kim, “Acoustic Ejector with Novel Lens Employing Air-
Reflectors,” the 19
th
IEEE International Conference on Micro Electro Mechanical
Systems, pp. 170-173, 2006
C.Y. Lee, S. Kamal-Bahl, H. Yu, J.W. Kwon, and E.S. Kim, “On-Demand DNA
Synthesis on Solid Surface by Four Directional Ejectors on a Chip,” IEEE/ASME
Journal of Microelectromechanical Systems, vol. 16, pp. 1130-1139, 2007.
C.Y. Lee, H. Yu, and E.S. Kim, “Nanoliter droplet coalescence in air by directional
acoustic ejection,” Applied Physics Letter, vol. 89, pp. 223902-223902-3, 2006
J. Lee, S.Y. Teh, A. Lee, H.H. Kim, C. Lee, and K.K. Shung, “Single beam acoustic
trapping,” Applied Physics Letter, vol. 95, pp. 073701-1-073701-3, 2009.
117
S.S. Lee, R.P. Ried, and R.M. White, “Piezoelectric Cantilever Microphone and
Microspeaker,” Journal of Microelectromechanical Systems, vol. 5, pp. 238-242,
1996.
H. Lintel, F. Pol, and S. Bouwstra, “A piezoelectric micropump based on
micromachining of silicon,” Sensors and Actuators, vol. 15, pp. 153-167, 1988.
M.
H. Ma, B. Hou, J. Gao, and C.Lin, “Development of One-sided Actuating
Piezoelectric Micropump Combined with Cold Plate in a Laptop,” Twenty-fourth
Annual IEEE Semiconductor Thermal Measurement and Management Symposium,
San Jose, CA, March 16-20, pp. 124-131, 2007.
P.L. Marston, “Axial radiation force of a Bessel beam on a sphere and direction
reversal of the force,” Journal of Acoustic Society of America, vol. 120, pp. 3518-
3524, 2006.
G.R. Matsueda and J.M. Stewart, “A p-Methylbenzhydrylamine Resin for Improved
Solid-Phase Synthesis of Peptide Amides,” Peptides, vol. 2, pp. 45-50, 1981.
R.B. Merrifield, “Solid Phase Peptide Synthesis. I. The Synthesis of a Tetrapeptide,”
Journal of the American Chemical Society, vol. 85, pp. 2149-2154, 1963.
G. Milne, G.D.M. Jeffries, and D.T. Chiu, “Tunable generation of Bessel beams with
a fluidic axicon,” Applied Physics Letters, vol. 92, pp. 261101-1-261101-3, 2008.
A.R. Mitchell, S.B.H. Kent, M. Engelhard, and R.B. Merrifield, “A New Synthetic
Route to tert-Butyloxycarbonylaminoacyl-4-(oxymethyl)phenylacet-amidomethyl-
resin, and Improved Support for Solid-Phase Peptide Synthesis,” Journal of Organic
Chemistry, vol. 43, pp. 2845-2852, 1978.
N.
K.C. Neuman, E.H. Chadd, G.F. Liou, K. Bergman, and S.M. Block,
“Characterization of Photodamage to Escherichia coli in Optical Traps,” Biophysical
Journal, vol. 77, pp. 2856-2863, 1999.
Q.
C. Qui, Q. Zhao, H. Zhang, W. Qu, H. Liu, and M. Cao, “A Valve-less PZT
Micropump with Isosceles Triangles Cross-section Diffuser Elements,” 1
st
IEEE
International Conference on Nano/Micro Engineered and Molecular Systems,
Zhuhai, China, January 18-21, pp. 200-203, 2006.
118
R.
M.B. Rasmussen, L.B. Oddershede, and H. Siegumfeldt, “Optical Tweezers Cause
Physiological Damage to Escherichia coli and Listeria Bacteria,” Applied and
Environmental Microbiology, vol. 74, pp. 2441-2446, 2008.
E. Riera, J.A. Gallego-Juarez, and T.J. Mason, “Airborne ultrasound for the
precipitation of smokes and powders and the destruction of foams,” Ultrasonics
Sonochemistry, vol. 13, pp. 107-116, 2006.
R.C. Roberts, J. Du, A.O. Ong, D. Li, C.Z. Zorman, and N.C. Tien, “Electrostatically
Driven Touch-Mode Poly-SiC Microspeaker,” Sensors 2007, IEEE Conference on
Sensors, Atlanta, GA, October 28-31, pp. 284-287, 2007
Y. Roichman, and D.G. Grier, “Projecting extended optical traps with shape-phase
holography,” Optics Letters, vol. 31, pp. 1675-1677, 2006.
S.
V.G. Sauerbrey, “Verwendung von Schwingquarzen zur Wägung dünner Schichten
und zur Mikrowägung”, Zeitschrift für Physik, vol. 155, pp. 206-222, 1959.
P. Sieber, “A new acid-labile anchor group for the solid-phase synthesis of C-
terminal peptide amides by the Fmoc method,” Tetrahedron Letters, vol. 28, pp.
2107-2111, 1987.
V. Singhal, “A novel valveless micropump with electrohydrodynamic enhancement
for high heat flux cooling,” IEEE Transactions on Advanced Packaging, vol. 28, pp.
216-230, 2005.
M. Shen and M.A.M Gijs, “High-performance magnetic active-valve micropump,”
Transducers ’09, International Solid-State Sensors, Actuators and Microsystems
Conference, Denver, CO, June 21-25, pp. 1281-1286, 2009.
T.
R. Tian-Ling, Z. Lin-Tao, L. Li-Tian, and L. Zhi-Jian, “Design Optimization of
Beam-Like Ferroelectrics-Silicon Microphone and Microspeaker,” IEEE
Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 49, pp.
266-270, 2002.
119
V.
V. Vivek, Y. Zeng, and E.S. Kim, “Novel Acoustic-Wave Micromixer,” Thirteenth
Annual International Conference on Micro Electro Mechanical Systems, pp. 668-673,
2000.
W.
S.S. Wang, “p-Alkoxybenzyl Alcohol Resin and p-Alkoxybenzyloxycarbonyl-
hydrazide Resin for Solid Phase Synthesis of Protected Peptide Fragments,” Journal
of American Chemical Society, vol. 95, pp. 1328-1333, 1973.
G. Whitworth, “Particle column formation in a stationary ultrasonic field,” Journal
of Acoustic Society of America, vol. 91, pp. 79-85, 1992.
J. Wu, “Acoustical Tweezers,” Journal of the Acoustical Society of America, vol. 89,
pp. 2140-2143, 1991.
Y.
H. Yu, J.W. Kwon, and E.S. Kim, “Microfluidic Mixer and Transporter Based on
PZT Self-Focusing Acoustic Transducers,” IEEE/ASME Journal of
Microelectromechanical Systems, vol. 15, pp. 1015-1024, 2006.
Abstract (if available)
Abstract
This thesis presents piezoelectric microelectromechanical systems (MEMS) for sound generation, liquid pumping, peptide synthesizing, and microparticle trapping. The common thread for the piezoelectric MEMS is the usage of piezoelectric lead zirconate titanate (PZT) sheet in generating sound waves for various applications in air and liquid. ❧ For air-borne audio sound generation, a microspeaker based on bending movement of PZT bimorph diaphragm was designed and fabricated. A PZT substrate was mechanically polished and glued to a micromachined silicon with low viscosity glue to minimize the glue thickness for minimum electric field drop in the glue. Top and bottom electrodes of the microspeaker were patterned according to the stress distribution of a diaphragm for uniform loading to optimize the piezoelectric actuation of bimorph structure. With an order of magnitude lower power consumption compared to a commercial cell-phone electromagnetic microspeaker, the newly fabricated microspeaker produced about 20 dB higher sound output in frequency range of 500 ~ 2,000 Hz. The power consumption increased as the frequency increased, but the sound output was still much higher than a commercial cell phone microspeaker. ❧ For liquid pumping, 260 μm-thick PZT sheet was actuated on its fundamental resonance frequency to generate 8.6 MHz acoustic waves. Acoustic Fresnel lenses were arranged along with U-shape and straight-line liquid channel to produce unidirectional liquid flow. The sector angle of pie-shape electrodes and Fresnel lens were optimized to generate larger in-plane direction acoustic body force, which directly affect the liquid flow rate in the channel. To improve the pumping rate and stabilize the operation condition, acrylic reflector was set up between Fresnel lenses. The micropump with U-shaped lens arrangement produced 7.3 cm/sec particle drift velocity on the liquid surface, and the micropump having straight-lined lens arranged and acrylic reflector showed pumping rate of 9.5 mL/min. ❧ For peptide synthesis using the Spot technique, acoustic droplet ejector with silicon lens structure was used to dispense pre-activated amino-acid on modified glass surface. Conventional acoustic ejector employing Fresnel lens with air-reflect uses parylene D polymer film as its structural material. The parylene acoustic Fresnel lens structure layer was replaced with bare silicon not only to simplify the fabrication steps but also to improve the durability of the device and convenience of washing out solution inside the ejector. With surface modified glass substrate and acoustic ejector as a liquid dispenser, 9-mers long peptide was synthesized with 70% of final synthesis yield that is equivalent with 96% stepwise synthesis yield. ❧ To expand the functionality of the acoustic droplet ejector as a liquid dispenser in peptide synthesizing system, an ejector array consisting of eight directional ejectors were designed and fabricated. The Fresnel lenses and electrodes of eight transducers were patterned into pie-shape with apex angle of 90° to reduce the distances between ejectors in an ejector array. The silicon wafer was etched with Deep Reactive Ion Etching (DRIE), and combined to the PZT transducer, with a mask aligner aiding on the alignment, to form ejecting chambers and liquid transfer channels. The chambers were designed to provide identical acoustic echo circumstances for each ejector in the array. The six ejectors out of the eight ejectors in an array were able to eject droplets onto a very small area of 238 x 380 μm² when actuated individually. ❧ For particle trapping, an acoustic tweezers was designed and fabricated on 128 &mu
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Piezoelectric ultrasonic and acoustic microelectromechanical systems (MEMS) for biomedical, manipulation, and actuation applications
PDF
Zero-power sensing and processing with piezoelectric resonators
PDF
High frequency ultrasonic phased array system and its applications
PDF
MEMS piezoelectric resonant microphone arrays and their applications
PDF
Array transducers for high frequency ultrasound imaging
PDF
High-frequency ultrasound array-based imaging system for biomedical applications
PDF
Acoustic ejector employing lens with air-reflectors and piezoelectrically actuated tunable capacitor
PDF
Development of novel 1-3 piezocomposites for low-crosstalk high frequency ultrasound array transducers
PDF
Wineglass mode resonators, their applications and study of their quality factor
PDF
Development of high frequency focused transducers for single beam acoustic tweezers
PDF
Highly integrated 2D ultrasonic arrays and electronics for modular large apertures
PDF
Ultrasonic microelectromechanical system for microfluidics, cancer therapeutics and sensing applications
PDF
Additive manufacturing of piezoelectric and composite for biomedical application
PDF
Development of front-end circuits for high frequency ultrasound system
PDF
Battery-less detection and recording of tamper activity along with wireless interrogation
PDF
Miniature phased-array transducer for colorectal tissue characterization during TEM robotic surgery; and, Forward-looking phased-array transducer for intravascular imaging
PDF
Surface acoustic wave waveguides for signal processing at radio frequencies
PDF
Design and development of ultrasonic array transducers for specialized applications
PDF
Design and implementation of frequency channelized ultra-wide-band (UWB) transceivers
PDF
2D ultrasonic transducer array’s design and fabrication with 3D printed interposer and applications
Asset Metadata
Creator
Choe, Youngki
(author)
Core Title
Audio and ultrasound MEMS built on PZT substrate for liquid and airborne applications
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Electrical Engineering
Publication Date
11/09/2012
Defense Date
10/17/2012
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
Fresnel lens,MEMS,OAI-PMH Harvest,piezoelectric,PZT,ultrasound
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Kim, Eun Sok (
committee chair
), Choma, John, Jr. (
committee member
), Shung, Koping Kirk (
committee member
)
Creator Email
ychoe@usc.edu,youngki.choe@gmail.com
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-107514
Unique identifier
UC11289538
Identifier
usctheses-c3-107514 (legacy record id)
Legacy Identifier
etd-ChoeYoungk-1276.pdf
Dmrecord
107514
Document Type
Dissertation
Rights
Choe, Youngki
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA
Tags
Fresnel lens
MEMS
piezoelectric
PZT
ultrasound