acoustofluidic manipulation of microspheres via vertical...
TRANSCRIPT
Master’s Thesis
Acoustofluidic Manipulation of Microspheres via Vertical
Hydrodynamic Focusing and Upward Migration
2017
Husnain Ahmed
Korea Advanced Institute of Science and Technology
Acoustofluidic Manipulation of Microspheres via Vertical
Hydrodynamic Focusing and Upward Migration
Husnain Ahmed
Advisor: Hyung Jin Sung
A dissertation/thesis submitted to the faculty of
Korea Advanced Institute of Science and Technology in
partial fulfillment of the requirements for the degree of
Master of Science in Mechanical Engineering
Daejeon, Korea
June 12, 2017
Approved by
Hyung Jin Sung
Professor of Mechanical Engineering Department
The study was conducted in accordance with Code of Research Ethics1.
1 Declaration of Ethical Conduct in Research: I, as a graduate student of KAIST, hereby declare that I have not
committed any acts that may damage the credibility of my research. These include, but are not limited to:
falsification, thesis written by someone else, and distortion of research findings or plagiarism. I affirm that my
thesis contains honest conclusions based on my own careful research under the guidance of my thesis advisor
MME
20154599
Abstract
Advancements in microfluidic manipulation devices are vital for the development of
future lab-on-a-chip technologies with applications in the biological, chemical, and materials
sciences. Surface acoustic wave (SAW)-based acoustofluidic separation devices, using low
power densities, offer non-contact and label-free sorting capabilities using differences in
mechanical properties (density, compressibility, etc.) or sizes of the micro-objects. A particle
suspended in a fluid within a microfluidic channel experiences a direct acoustic radiation force
(ARF) when travelling surface acoustic waves (TSAW) couple with the fluid at the Rayleigh
angle. Most SAW-based microfluidic devices rely on the horizontal component of the ARF to
migrate pre-focused particles laterally across a microchannel width. Although the magnitude
of the vertical component of the ARF is more than twice the magnitude of the horizontal
component, it has been long ignored due to polydimethylsiloxane (PDMS) microchannel
fabrication limitations and difficulties in particle focusing along the vertical direction. In the
present work, we have devised a single layered PDMS microfluidic chip for hydrodynamically
focusing particles in the vertical plane while explicitly taking advantage of the horizontal ARF
component to slow down the selected particles and the stronger vertical ARF component to
push the particles in the upward direction to realize continuous particle separation. An
acoustofluidic device with a straight PDMS microchannel placed directly on top of a straight
interdigitated transducer (IDT) was used to produce high frequency (140 MHz) TSAWs.
Contrary to the conventionally used two sheath flows for particle focusing, a single sheath flow
was used to pinch the particles close to the bottom of the microchannel. The TSAWs originating
from the IDT pushed the focused larger 4.8 µm diameter particles in the upward direction to
isolate them from smaller 2.0 µm or 3.2 µm diameter particles. The proposed particle
separation device offers high-throughput operation with purity > 97% and recovery rate > 99%.
It is simple in its fabrication and versatile due to the single layered microchannel design,
combined with vertical hydrodynamic focusing and the use of both the horizontal and vertical
components of the ARF. Using the concept of upward migration of particles, another
Husnain Ahmed. Acoustofluidic Manipulation of Microspheres via Vertical
Hydrodynamic Focusing and Upward Migration. School of Mechanical
Aerospace & Systems Engineering, Division of Mechanical Engineering.
2017. 38+vi pages. Advisor Prof. Sung, Hyung Jin.
acoustofluidic device was presented for the concentration and separation of four different sized
micro-objects inside a single-layered straight polydimethylsiloxane (PDMS) microchannel
without using external pumps. Two parallel placed interdigitated transducers (IDTs) were used
to produce high frequency (73 MHz & 140 MHz) traveling surface acoustic waves (tSAWs)
that trap and concentrate the 12 µm and 4.8 µm diameter particles at two different locations
inside the PDMS microchannel without the assistance of microfabricated PDMS membrane,
while allowing the 2.1 µm particles to filter through the chromatography of different size
microspheres.
Keywords: Acoustofluidics, particle separation, vertical component, acoustic radiation force,
trapping, concentration, chromatography.
i
Table of Contents
Table of Contents …………………………………………………………………………….……....i
List of Figures………………………………………………………………………………….……..ii
Chapter 1. Introduction
1.1. Acoustofluidic separation of vertically focused particles in a continuous flow………………….2
1.2. Acousto-chromatography based batchwise concentration and separation of particles…………...6
Chapter 2. Material and methods
2.1. Device fabrication………………………………………………………………………………...7
2.2. Tap based microchannel fabrication……………………………………………………………...9
2.3. Particles solution preparation……………………………………………………………………..9
2.4. Experimental setup………………………………………………………………………………..9
Chapter 3. Acoustofluidic separation of vertically focused particles in a continuous flow
3.1. Working Mechanism…………………………………………………….………………………11
3.2. Results and discussion……………………………………………….…………………………..15
3.2.1. Device operation at low flow rates…………………………………………………………….15
3.2.2. Characterization of inlets and outlets flow rate………………………………………………..17
3.2.3. Device operation at high flow rates……………………………………………………………19
3.2.4. Tape based microchannel for the separation of particles………………………………………23
3.3. Conclusions………………………………………………………………………………………23
Chapter 4. Acousto-chromatography based batchwise concentration and separation of particles
4.1. Working Mechanism………………………………………………………..……………………25
4.2. Results and discussion……………………………………………………….…………………..26
4.2.1. Concentration of particles by TSAWs…………………………………………………………26
4.2.2. Particles trapping from extremely low concentrated solution…………………………………30
4.2.3. A pumpless device for the concentration and separation of particles………………………….31
4.2.4. Naked eye view of device operation, sample collection and separation analysis……………...32
4.3. Conclusions………………………………………………………………………………………33
Bibliography…………………………………………………………………………………………35
Acknowledgement…………………………………………………………………………………...38
Curriculum Vitae…………………………………………………………………………………....39
ii
List of Figures
Figure 1.1. Tree diagram summarizes the most commonly used separation methods in
different categories. The current research route is shown in green……………………………..1
Figure 1.2. Uniformly spaced metal electrodes deposited on top of a piezoelectric substrate
(Lithium niobate, LiNbO3). Interdigitated transducer (IDT) was actuated by providing high
frequency AC signal. Only a certain AC frequency can actuate the IDT that depends on the
SAW speed 𝑐𝑆𝐴𝑊 , and inter-electrode gaps and widths
𝜆𝑆𝐴𝑊…………………………………..3
Figure 1.3. Calculation of the components of acoustic radiation force (ARF). The interaction
between TSAWs and the fluid results in leaky acoustic waves that radiate at an angle of ~22°
such that the vertical component Fv of ARF is more than twice its horizontal component
Fh…............................................................................................................................................4
Figure 1.4. Acoustofluidic particles separation using horizontal (a) and vertical (b)
components of acoustic radiation force (ARF)……………………………………………........5
Figure 2.1. (a) Deposition of Cr/Au metal electrodes on a LiNbO3 substrate by e-beam
evaporation method. (b) Fabrication of PDMS microchannel via soft lithography technique. (c)
Oxygen plasma bonding of PDMS microchannel with LiNbO3 substrate……………………...8
Figure 2.2. Experimental setup for acoustofluidics consisting of signal generator, DC power
supply, amplifier, micro syringe pump, microscope and camera, computer and acoustofluidic
device…………………………………………………………………………………………10
Figure 3.1. A schematic illustration showing the vertical migration of particles to realize size
based separation. The separation device is composed of a straight interdigitated transducer
(IDT) patterned on the lithium niobate (LiNbO3) substrate, a SiO2 layer and a straight PDMS
microchannel mounted on top. The top and side views of the device are illustrated as the
particles separation zone is enhanced when the power was turned off (a) and on (b),
respectively. Particles are slowed down by the horizontal component of the ARF Fh and pushed
in the upward direction depending on particle sizes due to the vertical component of ARF
Fh……………………………………………………………………………………………..13
iii
Figure 3.2. (a) A fabricated particle separation device. The PDMS microchannel was bonded
on top of the gold electrodes deposited onto a LiNbO3 substrate. Red dye was used to highlight
the microchannel and its ports. (b) The 3D solid geometry of the microchannel, with a top view.
(c) Side view of simulated streamlines within the straight microchannel……………………..14
Figure 3.3. (a) A 3D geometry and side view of the simulation of streamlines inside the straight
microchannel when the diameter of the punched hole is smaller than the width of the
microchannel. (b) The diameter of the punched holes is equal to the width of the
microchannel…………………………………………………………………………………15
Figure 3.4. (a) Simulation of the streamlines inside the microchannel at 𝑄1/𝑄2 = 𝑄4/𝑄3 with
zoom in side view of the streamlines after the 1st inlet. (b) Simulation of the streamlines inside
the microchannel at 𝑄1/𝑄2 < 𝑄4/𝑄3 with zoom in side view of the streamlines leaving
through 1st and 2nd outlet……………………………………………………………………...15
Figure 3.5. Photographic images of the particle separation experiment based on the upward
movement of particles under an applied ARF. (a) Power off: the particle mixture (green & red)
flowed together through the lower streamlines and resulted in no separation. (b) Power on:
green particles migrated upward toward the upper streamlines, resulting in separation………16
Figure 3.6. A side view graphic supporting the photographic images at the outlet pipes. (a)
Power off: a control experiment showing that particles of different sizes were collected through
the 2nd outlet. (b) Power on: Larger (green) and smaller (red) particles were collected through
1st and 2nd outlet, respectively.................................................................................................17
Figure 3.7. Particle separation results for 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1. Flow cytometry
graphs showing the ratio of particles of different sizes collected at the collection (a) and waste
(b), outlets respectively……………………………………………………………………….18
Figure 3.8. Particle separation purity measures at different ratios of the inlet and outlet flow
rates. (a) Percentage of purity at various outlet flow ratios 𝑄3: 𝑄4 was varied from 9:1 to 0.67:1,
holding the inlet flow ratio fixed at 𝑄1: 𝑄2 = 1: 9. (b) Percentage of purities across inlet flow
ratios of 𝑄1: 𝑄2 = 1: 9 to 1: 1, holding the outlet flow ratio fixed at 𝑄3: 𝑄4 = 0.67: 1. (c), (d)
Hemocytometer images obtained showing the particle size compositions collected at different
outlet and inlet flow ratios.........................................................................................................19
iv
Figure 3.9. (a) A side view schematic diagram illustrating the height of the sheath flow and the
mixed particle flow within sections of the microchannel. Particles with 𝜅 > 1 were affected by
the ARF, whereas particles with 𝜅 < 1 were predominantly influenced by the streaming flow.
(b) Top view of the experimental images, illustrating the effects of SAWs on particles of
different sizes…………………………………………………………………………………20
Figure 3.10. Purity and recovery measures at different net flow rates QNET by holding the
conditions (𝑄1/𝑄2 = 1/9) and (𝑄3 < 𝑄4) for the isolation of green, 4.8 µm and red, 3.2 µm
particles………………………………………………………………………………………21
Figure 3.11. Hemocytometer images of sample collection at first outlet. Varied the net
flowrates (𝑄NET) by fixing the inlets and outlets flow ratio, 𝑄1 : 𝑄2 = 1 : 9 and 𝑄3 < 𝑄4 for
the separation of green, 4.8 µm and red, 3.2 µm particles………………………...…………...22
Figure 3.12. Hemocytometer images of sample collection at second outlet. Varied the net
flowrates (𝑄NET) by fixing the inlets and outlets flow ratio, 𝑄1 : 𝑄2 = 1 : 9 and 𝑄3 < 𝑄4 for
the separation of green, 4.8 µm and red, 3.2 µm particles……………………………………..22
Figure 3.13. Simulation of the streamlines inside 250 µm wide microchannel with 1mm
punched hole at 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1 for different net flow rates…………..23
Figure 3.14. Simulation of the streamlines inside 250 µm wide microchannel with 500 µm
punched hole at 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1 for different net flow rates……………23
Figure 3.15. (a) A top view and (b) side view of the schematic with experimental images of
the manually fabricated scotch tape based microchannel……………………………………..24
Figure 4.1. A schematic diagram showing the concentration and separation of three different
diameter particles. The device is composed of two parallel placed straight interdigitated
transducer (IDT) patterned on the lithium niobate (LiNbO3) substrate, a SiO2 layer and a
straight PDMS microchannel loosely positioned on top without plasma bonding. The top and
side views of the device are showed as the concentration and separation zone is enlarged during
SAW off (a) and on (b) respectively. Particles slowed down by the horizontal component of
v
ARF Fh and concentrated by the vertical force component Fv depending on particle
diameters………………………………………………………………………………..........27
Figure 4.2. A schematic diagram supported by a photographic image of particle concentration
inside the microchannel due to the combination of horizontal and vertical force components.
Hemocytometer images of inlet sample solution and outlet collected sample illustrated the
efficient concentration of microspheres inside the microfluidic channel……………………..28
Figure 4.3. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 400 particles/µL. (a) Experimental images of top view
of the microchannel and graph (b) showing the linear increase in the particle concentration
from 1 sec to 30 sec. (c) Bar chart illustrate the decrease in the number of particles per microliter
in the sample solution results in the increase in required time to achieve fixed
concentration…………………………………………………………………………………28
Figure 4.4. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 200 particles/µL. (a) Experimental images of top view
of the microchannel and graph (b) showing the linear increase in the particle concentration
from 6 sec to 60 sec…………………………………………………………………………...29
Figure 4.5. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 100 particles/µL. (a) Experimental images of top view
of the microchannel and graph (b) showing the linear increase in the particle concentration
from 15 sec to 120 sec………………………………………………………………………...29
Figure 4.6. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 20 particles/µL. (a) Experimental images of top view
of the microchannel and graph (b) showing the linear increase in the particle concentration
from 1 min to 10 min………………………………………………………………………….30
Figure 4.7. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 10 particles/µL. (a) Experimental images of top view
of the microchannel and graph (b) showing the linear increase in the particle concentration
from 3min to 30 min…………………………………………………………………………..30
vi
Figure 4.8. Characterization of particle concentration inside the microchannel. The
concentration of the sample solution is 1 particle/µL. (a) Experimental images of top view of
the microchannel and graph (b) showing the linear increase in the particle concentration from
10 min to 120 min…………………………………………………………………………….31
Figure 4.9. Efficient particles trapping by the vertical component of ARF from extremely low
concentrated particles solution. The concentration of sample solution is 10 particles/ml. 99 out
of 100 targeted particles were captured in 2 hour……………………………………………..32
Figure 4.10. Photographic images of the particle concentration and separation experiment
based on the stronger force component of an applied ARF imposed by TSAWs. (a) Power on:
blue particles passed through the chromatography of red and green particles. (b) Power on:
after washing with ethanol red and green particles concentration is displayed. (c) Naked eye
view of the concentrated (different diameter) particles inside the microchannel. (d)
Hemocytometer images obtained displaying the particle size compositions of concentrated and
separated particles…………………………………………………………………………….33
Figure 4.11. Photographic image of the experimental setup for naked eye view of the particles
chromatography and sample collection from the outlet for separation analysis………………34
1
Chapter 1. Introduction
Microfluidic particle and cell separation takes over the conventional method of cell
separation. The later method usually employed membrane based filtering schemes which are
limited by the membrane pore size and are easily susceptible to clogging after a particular time
period. In contrast, microfluidics offer various advantages including: reduced sample volumes,
fast sample processing, high sensitivity, low cost, integrated systems minimized the chances of
sample contamination and increased flexibility in potential for point-of-care diagnostic in the
areas which lacks clinical labs and skilled staffs. In the competition to develop innovative
microfluidic tools, acoustofluidics is competing with other microfluidic technologies like,
opto-fluidics, dielectrphoresis, electrophoresis, hydrodynamic filtration and inertial
microfluidics. A summary of most commonly used particle or cell separation methods along
with their classification is shown in the tree diagram (see Figure 1.1).
Figure 1.1. Tree diagram summarizes the most commonly used separation methods in different categories. The
current research route is shown in green.
In the present work, we designed two acoustofluidic devices for the continuous (device
1) and batch-wise separation (device 2) of different diameter particles. In the following
Label based
Label free
Current research
route
Active methods
Cell/particle separation methods
Acoustofluidics
Dielectrphoresis
Electrophoresis
Optical sorting
FACS
MACS
Hydrodynamic filtration
Pinched flow fractionation
Inertial microfluidics
Pillar and weir structure
SAW
BAW
TSAW
SSAW
Horizontal force component
Vertical force component
Active techniques: rely on an external force field for functionality.
Passive techniques: rely entirely on the channel geometry and inherent hydrodynamic forces for functionality.
Label based techniques: the cell surface markers are identified via specific antibodies labeled with fluorescent molecules or magnetic beads. e.g. fluorescence-activated cell sorting (FACS) & magnetically actuated cell sorter (MACS)
Label free techniques: separation is often based on intrinsicproperties of the sample components, such as their size, charge or their polarizability.
SAW – Surface Acoustic WavesBAW – Bulk Acoustic WavesTSAW – Travelling Surface Acoustic WavesSSAW – Standing Surface Acoustic Waves
Passive methods
Separation markers:
shape, size, density, compressibility
2
subsections, we briefly summarized previous literature related to the separation and
concertation of particles and proposed our new findings.
1.1. Acoustofluidic separation of vertically focused particles in a
continuous flow
Over the past few decades, microfluidic particle and cell separation techniques have
played a vital role in cell biology, disease diagnostics, drug screening, drug discovery, and
biochemical analysis.1,2 For example, cancer treatments requires the isolation of circulating
tumor cells (CTCs) from peripheral blood samples.3,4 Likewise, the cure for malaria relies on
the separation of malaria-infected red blood cells from healthy cells.5–7 Sorting the different
types of stem cells, such as embryonic stem cells (hESCs),8,9 tissue-specific stem cells,10
mesenchymal stem cells (MSCs)11 and induced pluripotent stem cells (iPSCs)12,13 is essential
for the analysis of different diseases and organ-on-a-chip technology14 for the development of
new therapies and drugs.
To date, many researchers have explored various microfluidic methods for the
manipulation of particles inside minute volumes of fluids, such as inertial microfluidics,15
hydrodynamic filtration,16 magnetophoresis,17 dielectrophoresis,18,19 optofluidics,20,21 and
acoustophoresis.22–24 Acoustophoresis-based microfluidic separation techniques have been
preferred due to the contactless handling of the biological samples, low power requirement,
biocompatible nature of acoustic waves, etc. that ease their incorporation into micro total
analysis systems. Acoustophoresis based devices are classified based on their driving
mechanisms: bulk acoustic wave (BAW) and surface acoustic wave (SAW). BAW devices
consist of a piezoelectric transducer attached to microfluidic chip with acoustic waves
propagating in bulk of the material25 which generates ultrasonic standing waves to concentrate
or separate the particles inside the microchannel. However, in SAW based particle separation
devices a pair of interdigitated electrodes patterned on a piezoelectric substrate (see Figure
1.2) produces acoustic waves that travel along the surface of the substrate to manipulate the
particles.26 As most of the acoustic energy is concentrated on the surface of the substrate, SAW-
based devices are more energy efficient.
SAWs are classified into standing surface acoustic waves (SSAWs) and travelling
surface acoustic waves (TSAWs).27 A SSAW is a combination of two constructively interfering
3
TSAWs propagating in opposite directions. SSAWs form pressure nodes and anti-nodes inside
the fluid and are used to push particles toward regions of low pressure inside the microchannel,
achieving particle focusing and separation.28–30 On the other hand, TSAWs have been used in
Figure 1.2. Uniformly spaced metal electrodes deposited on top of a piezoelectric substrate (Lithium niobate,
LiNbO3). Interdigitated transducer (IDT) was actuated by providing high frequency AC signal. Only a certain AC
frequency can actuate the IDT that depends on the SAW speed 𝑐𝑆𝐴𝑊, and inter-electrode gaps and widths 𝜆𝑆𝐴𝑊.
cross-type acoustic particle separators to laterally migrate particles and realize separation
across the microchannel width or within a sessile droplet, because particles predominantly
migrate within the horizontal plane.31–34 Most SAW-based acoustofluidic separation techniques
utilize forces that act on micro-objects suspended in a horizontal plane while pushing them
laterally inside the microchannel.35–39 The interaction between TSAWs and the fluid results in
leaky acoustic waves that radiate at an angle of ~22° (in systems comprising water and a lithium
niobate substrate) inside the microfluidic channel, such that the vertical component (Fv) of the
acoustic radiation force (ARF) acting on the suspended particles is ~2.5 times greater than the
horizontal component of the force (Fh), i.e., Fv ≅ 2.5Fh (see Figure 1.3). The SAW-based
acoustofluidic devices that utilize the horizontal component of an ARF are composed of an
interdigitated transducer (IDT) integrated into the side of a single layered PDMS
microchannel37,40,41 (see Figure 1.4(a)). A multiple-layered PDMS microchannel with a post
beneath the micro-object manipulation zone was also used in a similar fashion to deflect
particles37 and sort cells or droplets.35 IDTs may also be positioned directly beneath a
microchannel to induce desired particle migration. Collins et al. used a single IDT to produce
standing acoustic waves within a single layered microchannel to achieve particle separation.42
However, TSAWs, produced by an IDT placed beneath a single layered microchannel, have
not yet been demonstrated to be capable of inducing vertical particle migration by imposing a
λλ/4
SAW SAW
Straight Interdigitated transducer (SIDT)
LiNbO3
~
𝑆𝐴𝑊 =𝑐𝑆𝐴𝑊𝜆𝑆𝐴𝑊
= 𝐴
4
direct ARF onto particles directly from the bottom of the microchannel (see Figure 1.4(b)).
Previously our research group developed a cross-type particle separation device that showed
vertical particle migration, although the effect was not harnessed for an application.24 Collins
et al.43 took advantage of the vertical component of the ARF and employed a focused IDT to
trap and concentrate selected particles behind a micro-fabricated PDMS membrane. However,
the particle trapping capacity of the membrane would reach a saturation value in a short period
of time; therefore, a continuous flow separation device offers improved particle separation
performance..22,29
Figure 1.3. Calculation of the components of acoustic radiation force (ARF). The interaction between TSAWs
and the fluid results in leaky acoustic waves that radiate at an angle of ~22° such that the vertical component Fv
of ARF is more than twice its horizontal component Fh.
In the present work, a straight IDT was deployed with a straight loosely aligned
microchannel positioned perpendicularly to the uniformly spaced metal electrodes. Compared
to most SSAW-based devices29,44,45 that usually require a pair of parallel IDTs tightly aligned
with the microchannel, the present device did not require tight alignment with the microchannel
similar to other TSAW-based devices.22,24 However, tilted angle SSAWs have been utilized to
circumvent such limitations.46 The proposed design provided an additional advantage of
utilizing both the horizontal and stronger vertical components of the ARF acting on the
particles. The height and width of the microchannel did not significantly alter the device
performance, unlike the SSAW-based devices, in which the microchannel aspect ratio
significantly affected the locations of the pressure nodes and anti-nodes.28,44,45,47 The present
SAW
x
Y+ + --
Side viewLiNbO3
Snell’s law:
= 90
= 𝑐
𝑐 ( )
TSAW
LeakyTSAW
𝑐
𝑐
𝑐 000
𝑐 1 00
=
;
= = .
5
device utilized a comparatively low input power because the energy loss to PDMS walls was
minimal as the IDT was directly exposed to the fluid and both the vertical and horizontal
components of the ARF were employed. A single layered PDMS microchannel with a simple
design was used to continuously isolate the selected particles based on their diameters by
inducing vertical migration. Furthermore, particle focusing based on two sheath flows, as used
in most SAW-based particle separation devices,4,23,29,38 was circumvented in the present work
by focusing the particles along the vertical direction using a single sheath flow. Note that some
SPLITT or H-filter based separation devices used a single sheath flow configuration. For
instance, Hawkes et al.48 used a single sheath flow to focus the yeast cells along the side of the
microchannel wall prior to washing them using SSAWs. However, the migration of the cells
was in horizontal direction and the device required a fabrication of multiple layered
microchannel. Recently Chen et al.49 demonstrated the separation of platelets from whole blood
by focusing the biological sample vertically and then pushing them upward using BAWs in a
multiple layered microchannel. In contrast, the present device strictly distinguishes the external
force components (horizontal from vertical), which is available with TSAW device only. The
mechanism for SSAW and BAW based devices would be different. Among the single layered
PDMS based devices, similar kind of focusing is not reported before. Due to the utilization of
the principal component of ARF, the present device successfully operated for particle
separation at net flow rate up to 1.3 mL/min which is approximately 100 times higher than used
in previously reported SAW devices.22–24,29,38,39,42,47,50,51 The throughput of the proposed device
can be further improved with ease by increasing the width of the microchannel since the
microchannel is placed directly on the IDT. This low power, high throughput tiny device will
be useful for point-of-care testing applications based on biological sample separation.
Figure 1.4. Acoustofluidic particles separation using horizontal (a) and vertical (b) components of acoustic
radiation force (ARF).
TSAW in
horizontal
direction
Top view
x
z
SAW
IDT(a)
SAW
TSAW in
vertical
direction
x
Y
Side view PDMS microchannel
(b)
6
1.2. Acousto-chromatography based batchwise concentration and
separation of particles
The separation and concentration of a specific analyte such as circulating tumor cells or
malaria-infected red blood cells are fundamental for its diagnostic or therapeutic efficacy. In
addition to the high throughput separation of particles in a continuous flow, a pumpless
microfluidic device was designed to trap and concentrate the extremely low concentrated
particles solution. As described earlier, Collins et al.43 employed the focused IDT to
concentrate the selected particles behind the microfabricated membrane inside the
microchannel, while allowing the smaller particles to separate. However, the fabrication of
complex geometry of microchannel and micro-membrane is difficult in comparison to the
fabrication of a single layered PDMS microchannel. Li et al concentrated the particles inside
the sessile droplet of water by using SAWs. Similarly Destgeer et al. concentrated and
separated the different sized particles inside the sessile droplet.31 However, this kind of
separation can only be useful for diagnostics which requires only visualization. It’s not possible
to collect the sample after separation of particles inside the sessile droplet; as, the distinct
regimes of different sized particles might get disturb while collecting the separated samples
with pipette.
In the present work, a SAW-based pump-free particle concentration and separation
device was proposed, which used a single-layer PDMS microchannel (without microfabricated
membrane) to concentrate the particles based on their diameter at various locations inside the
microchannel by taking advantage from the vertical component of ARF, while allowing the
smaller particles to separate through the concentration of different size micro-particles. After
concentration of particles inside the microchannel we successfully collected each separated
sample at the outlet for analysis.
7
Chapter 2. Materials and methods
The fabrication of acoustofluidic devices was carried out in two steps. At first step metal
electrodes were deposited on a piezoelectric substrate while in the second step PDMS
microfluidic channel was fabricated using soft lithography technique (see Figure 2.1). These
fabrication processes and particles solution preparation is further discussed in detail in the
following subsections.
2.1. Device Fabrication
The acoustofluidic device as shown in Figure 3.2(a) was composed of a piezoelectric
(Lithium Niobate, LiNbO3) substrate (500 µm thick, 128 Y-X cut, MTI Korea, Korea) with
an IDT deposited on top and a PDMS microfluidic channel. A bimetallic (Cr/Au, 300 Å/1000
Å) layer of interdigitated electrodes was deposited, using the e-beam evaporation method and
lift-off process, to form an IDT having a total aperture of 1 mm and 20 electrode finger pairs
with uniform widths and spacings in between (λ/4 = 6.5 μm). A thin layer of SiO2 (2000 Å)
was also deposited on top of the electrodes using plasma enhanced chemical vapor deposition
to keep them safe from mechanical damage and enhance bonding. The whole process is brifely
explained in Figure 2.1(a).
A commonly used soft photolithographic process was used to fabricate PDMS
microchannel; 80 µm thick layer of SU-8 2100 (Micro Chem, USA) was spin coated on top of
a Si substrate. Later on, after soft baking at 65 °C for 5 min and 95 °C for 20 min photoresist
was exposed to UV light through hard contact of patterned chrome mask to make the desired
shape of the microchannel. After that, the SU-8 pattern was post-baked at 65 °C for 5 min and
95 °C for 10 min. At the end, it was dipped in the SU-8 developer to remove the unexposed
area of negative photoresist. Finally, the desired SU-8 pattern was rinsed with isopropyl alcohol
and DI water. The SU-8 pattern on Si substrate acted as a mold that could be transmitted to the
PDMS. A PDMS base and a curing agent (Sylgard 184A & 185B, Dow corning, U.S.A) mixed
in a 10:1 ratio and poured on top of the SU-8 mold. The bubbles formed during mixing of the
PDMS base and the curing agent were removed by putting the mixture under vacuum. PDMS
was cured at 65 °C for at least two hours and later on peeled off the Si substrate. SU-8 mold
8
was exposed to trichloro (1H, 1H, 2H, 2H-perfluorooctyl) silane (Sigma-Aldrich, U.S.A) under
vacuum for the surface treatment prior to pouring PDMS mixture. It helps in the easy release
of PDMS from the SU-8 mold after hard baking. After the inlet and outlet ports were punched
through the PDMS using punching tool (Harris Uni-Core), the microchannel and LiNbO3
substrate were exposed to oxygen plasma52 in a multipurpose oxygen plasma system (Covance,
Femto Science, Korea) for 2 min at 150 W and 750 mTorr. The microchannel was aligned
perpendicular to the IDT and bonded to the substrate in such a way that IDT would be manually
placed between the second inlet and first outlet. As the PDMS microchannel was 250 µm in
width and 80 µm in height, it did not require a tight alignment against the IDT with 1 mm
aperture. This process of microchannel fabrication is brifely explained in Figure 2.1(b,c).
Figure 2.1 (a) Deposition of Cr/Au metal electrodes on a LiNbO3 substrate by e-beam evaporation method. (b)
Fabrication of PDMS microchannel via soft lithography technique. (c) Oxygen plasma bonding of PDMS
microchannel with LiNbO3 substrate.
LiNbO3 substrate
Positive photoresist
UV exposure through chrome mask
Develop photoresist
Deposit Cr/Au layer
Lift –off process
(a) Electrode deposition on LiNbO3
Si substrate
Negative (SU-8) photoresist
UV exposure
Develop photoresist
Pour PDMS and cure
Peel off PDMS
(b) Fabrication of PDMS microchannel
(c) Oxygen plasma bonding
SAW device PDMS microchannel
Final device
Deposition of SiO2 layer
9
2.2. Tape based microchannel fabrication
In addition to standard MEMS and soft lithography method of microchannel fabrication,
we also fabricated the microchannel using scotch tape manually; as height and width of the
microchannel does not play a critical role in our device for particle separation. In contrast to
spin coating a photoresist on top of a silicone substrate and then making a mold of a photoresist
using photomask and UV aligner, a piece of a polyimide silicone adhesive tape (thickness of
70 µm that defines microchannel height, Tianjin Fortune, China) was manually cut (width ~800
µm) and pasted on top of the Si substrate which acted as a mold for the PDMS microchannel.
After a mixture of PDMS and its curing agent (10:1) was poured on top of the scotch-tape mold,
it was vacuumed to remove unnecessary bubbles, and was baked at 65 °C for at least two hours.
Rest of the process is similar to what has already been explained in the previous section. This
process of making a microchannel could be a cost-effective alternative compared to the
standard soft lithography process.
2.3. Particles solution preparation
Three different sized polystyrene particles having the same density were used in two pairs
(4.8 µm + 2.0 µm, and 4.8 µm + 3.2 µm) to demonstrate the separation mechanism: 4.8 µm
green fluorescent (<5% uniformity, No.G1000, Thermo Scientific, CA, USA), 3.2 µm red
fluorescent (<5% uniformity, No.R0300, Thermo Scientific, CA, USA) and 2.0 µm red
fluorescent (<5% uniformity, No.R0200, Thermo Scientific, CA, USA). Each of the particle
solution consisted of 1% solid microspheres per microliter of liquid as the total number of
beads/µl were counted (green 4.8 µm: 1.72 × 105, red 3.2 µm: 5.81 × 105, red 2.0 µm: 2.39 ×
106) for each particle solution. To add an equal number of particles in a sample, 13.9 µL of
green (4.8 µm) and 1 µL or 3.4 µL of red (2.0 µm or 3.2 µm) particle solutions per 1 ml of DI
water were added. Later on, 21.8 wt. % glycerin and 1 wt. % tween 20 was added.53 The
addition of glycerin matched the density of the particles with water and prevented the particles
from settling down in water. Tween 20 was added to avoid the formation of doublets and
triplets of the particles by preventing particle-to-particle adhesion.
10
2.4. Experimental setup
A schematic of experiment layout is shown in Figure 2.2. The device was mounted on
top of a fluorescent microscope stage (BX53, Olympus, Japan) for visualization using an
appropriate lens (2x, 4x or 10x magnification). The mixture of fluorescent polystyrene particles
with diameters 4.8 µm (green) and 2.0 µm (red) was pumped (neMESYS, Cetoni GmbH,
Germany) into the microchannel through the first inlet. DI water was pumped as a sheath flow
through the second inlet to force the particles solution to flow in the lower streamlines. One of
the two outlets was connected with a pump to suck the fluid out by applying a negative pressure
and another outlet was opened to atmospheric pressure for the collection of sample for analysis.
The inlet flow rates used were 50 µL/hr and 450 µL/hr for sample and sheath flows,
respectively, unless otherwise mentioned.
Figure 2.2 Experimental setup for acoustofluidics consisting of signal generator, DC power supply, amplifier,
micro syringe pump, microscope and camera, computer and acoustofluidic device.
The flow rates at the outlets were readily adjusted to achieve a maximum separation. An RF
signal generator (N5181A, Agilent Technologies, U.S.A) was used to generate an AC signal at
140 MHz frequency and tunable amplitude 4.5-50 mV that was amplified (LZT-22+, Mini-
Circuits, U.S.A) up to 11-1330 mW before feeding into the device. The images of green and
red fluorescent particles were captured by CCD camera (DP72, Olympus, Japan) using separate
filters and later on stacked together to obtain a single image. The captured microscopic images
Signal generator, N5181A [3 GHz]
DC power supply, E3634A
μ-pump, neMESYS
Microscope, BX51
Camera, DP72
Power amplifier,
ZHL-1-2W
1st
Inlet
2nd
Inlet
1st
outlet
2nd
outlet
LiNbO3
Au electrodes
PDMS
11
were processed to remove the background noise using ImageJ (http://imagej.nih.gov/ij/)
software for better image quality. The collected sample particles at two outlets were analyzed
through ImageJ software and counter-analyzed with a c-chip disposable hemocytometer
(Digital Bio, Korea). At the end, the particles were counted using flow cytometry at the specific
flow ratios at the inlets and outlets to confirm the efficient separation of the particles.
12
Chapter 3. Acoustofluidic separation of vertically focused
particles in a continuous flow
3.1. Working Mechanism
A schematic diagram of the particle separation device displays a simple PDMS
microchannel attached to a piezoelectric substrate (LiNbO3) with an IDT deposited on top of it
(see Figure 3.1). A silicon dioxide (SiO2) layer on top of the IDT protected the electrodes from
mechanical damage and enhanced PDMS–substrate bonding. A straight PDMS microfluidic
channel consisting of two inlet ports and two outlet ports was mounted on top of the IDT in
such a way that the IDT was positioned between the second inlet and the first outlet. A mixture
of two different sized particles (larger green, smaller red) was injected through the first of the
two inlets at a flow rate of 𝑄1, and a sheath fluid in the form of deionized (DI) water was
introduced through the second inlet at a flow rate of 𝑄2. The purpose of the sheath flow was to
pinch the sample mixture in the lower streamlines using the DI water flowing in the upper
streamlines to create a vertical double-layered flow while hydrodynamically focusing the
particles close to the bottom of the microchannel.54,55 The first of the two outlets was used to
pump out the fluid at a flow rate of 𝑄3 by applying a negative pressure, while the remaining
fluid was collected through the second outlet, which was open to the atmosphere, with a flow
rate of 𝑄4. The sheath flow induced all particles to flow through the lower fluidic streamlines,
and particles could be collected through the second outlet, without separation when the power
was off (see Figure 3.1(a)). Previous studies demonstrated that the value of the 𝜅-factor,
defined as 𝜅 = 𝜋𝑑 /𝑐 , where 𝑑 is the particle diameter, is the TSAW frequency, and 𝑐 is
the speed of sound in the fluid, could be used to effectively characterize the particle motion
under a TSAW.27 The TSAW frequency and the particle diameters were chosen such that 𝜅 >
1 for larger (green) particles and 𝜅 < 1 for smaller (red) particles. Once the AC signal was
applied to the IDT, it generated an acoustic wave that applied a significant ARF to the larger
(green) particles in the horizontal and vertical directions. The horizontal (X-direction)
component of the ARF (Fh) slowed down the particles against the direction of the flow due to
the resultant drag force Fd on the particle, whereas the major (Y-direction) component of the
ARF (Fv) pushed the green particles vertically upward into the upper streamlines, which were
ultimately collected through the first outlet. On the other hand, smaller (red) particles continued
to flow in the lower streamlines, nearly unaffected by the ARF, and were ultimately collected
13
through the second outlet (see Figure 3.1(b)). As a result, the green particles and red particles
were continuously separated through the first and second outlets based on their size difference.
Figure 3.1. A schematic illustration showing the vertical migration of particles to realize size based separation.
The separation device is composed of a straight interdigitated transducer (IDT) patterned on the lithium niobate
(LiNbO3) substrate, a SiO2 layer and a straight PDMS microchannel mounted on top. The top and side views of
the device are illustrated as the particles separation zone is enhanced when the power was turned off (a) and on
(b), respectively. Particles are slowed down by the horizontal component of the ARF Fh and pushed in the upward
direction depending on particle sizes due to the vertical component of ARF Fh.
A fabricated particle separation device is shown in Figure 3.2(a). A red dye was used to
highlight the microchannel, along with the inlet and outlet ports. The vertical migration of
particles induced by the stronger Y-directional component of the ARF could be harnessed to
induce particle separation only in the presence of the double-layered flow within the
microchannel. The streamlines in the microchannel were simulated using various inlet and
outlet port sizes to optimize the geometry for stabilizing the double-layer flow. The 3D
geometry of the microchannel is shown in Figure 3.2(b). If the diameter of the inlets and outlets
holes were smaller than the width of the microchannel, the streamlines from the first inlet
followed a path around the streamlines coming from the second inlet. This flow resulted in a
horizontally triple-layered flow that was not suitable for separating particles, because they
weren’t focused prior to exposure to the TSAW (see Figure 3.3(a)). If the size of the punched
14
holes was equal to (see Figure 3.3(b)) or larger than the width of the microchannel, however,
the device favored the desired double-layered flow, as the streamlines (blue) emerging from
the second inlet readily pinched the streamlines (red) emerging from the first inlet (see Figure
3.2(c)). The streamlines originating from the first and second inlets were collected through the
second outlet and first outlet, respectively. In general, it was not possible to manually punch
inlet/outlet holes with a diameter equal to the width of the microchannel; therefore, a
microchannel design with inlet and outlet ports larger than the microchannel width was
preferred. A top view of the microchannel inlet port, shown in Figure 3.2(b), reveals that the
diameter D of the punched hole exceeded the width w of the microchannel to form a double-
decker flow inside the microchannel, as predicted by the simulation results shown in Figure
3.2(c). The breadth of the punched holes (with respect to the width of the microchannel) was
fixed, and the streamlines at different flow ratios of the inlets and outlets were simulated. The
purpose of this simulation was to observe the height ratio of the sheath flow streamlines and
particle mixture streamlines during steep focusing of the sample mixture streamlines in the
perpendicular plane (see Figure 3.4(a)), and to study the consequences of the streamline height
(at the outlets) on the separation efficiency (see Figure 3.4(b)).
Figure 3.2. (a) A fabricated particle separation device. The PDMS microchannel was bonded on top of the gold
electrodes deposited onto a LiNbO3 substrate. Red dye was used to highlight the microchannel and its ports. (b)
The 3D solid geometry of the microchannel, with a top view. (c) Side view of simulated streamlines within the
straight microchannel.
15
Figure 3.3. (a) A 3D geometry and side view of the simulation of streamlines inside the straight microchannel
when the diameter of the punched hole is smaller than the width of the microchannel. (b) The diameter of the
punched holes is equal to the width of the microchannel.
Figure 3.4. (a) Simulation of the streamlines inside the microchannel at 𝑄1/𝑄2 = 𝑄4/𝑄3 with zoom in side view
of the streamlines after the 1st inlet. (b) Simulation of the streamlines inside the microchannel at 𝑄1/𝑄2 < 𝑄4/𝑄3
with zoom in side view of the streamlines leaving through 1st and 2nd outlet.
16
Figure 3.5. Photographic images of the particle separation experiment based on the upward movement of particles
under an applied ARF. (a) Power off: the particle mixture (green & red) flowed together through the lower
streamlines and resulted in no separation. (b) Power on: green particles migrated upward toward the upper
streamlines, resulting in separation.
3.2. Results and discussion
3.2.1. Device operation at low flow rates
A mixture of microspheres was separated using a single IDT placed under a straight
microchannel, as shown in Figure 3.5. Experimental images revealed the extent of particle
separation under the applied vertical component of the ARF. The particle mixture (green, 4.8
µm & red, 2.0 µm) and DI water were injected through the first and second inlets with flow
rates of 50 µL/hr (𝑄1) and 450 µL/hr (𝑄2), respectively. The first and second outlets were used
to collect the particles at flow rates of 200 µL/hr (𝑄3) and 300 µL/hr (𝑄4), respectively. When
the device was turned off, the large (green) and small (red) particles flowed together over the
IDT along the lower streamlines below the sheath flow streamlines, as shown in Figure 3.5(a).
These conditions resulted in no separation of particles. The 𝜅-factor values for the 4.8 µm
particles (green) and 2.0 µm particles (red) were calculated to be 1.407 and 0.586 ,
respectively.23 Therefore, when the power was turned on, the green particles were pushed
17
upward due to the vertical component of the ARF, because 𝜅 > 1, whereas the smaller red
particles with 𝜅 < 1 flowed along the lower streamlines, as shown in Figure 3.5(b). The green
particles were collected at the first outlet whereas the red particles moved toward the second
outlet. These effects produced the successful separation of particles, as indicated in the
photographic images of the outlets, shown in Figure 3.6.
Figure 3.6. A side view graphic supporting the photographic images at the outlet pipes. (a) Power off: a control
experiment showing that particles of different sizes were collected through the 2nd outlet. (b) Power on: Larger
(green) and smaller (red) particles were collected through 1st and 2nd outlet, respectively.
A side view of the outlet pipes confirmed the separation of particles. When the power
was turned off, all of the particles (green & red) were collected at the second outlet. Not a
single particle flowed through the first outlet, as shown in Figure 3.6(a). When the power was
turned on, however, the particles were effectively separated as the larger particles (green)
passed through the first outlet and the smaller (red) particles passed through the second outlet,
as shown in Figure 3.6(b).
Flow cytometry was used to count the particles collected from the first outlet (collection)
and the second outlet (waste) under the experimental conditions 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 =
0.67/1, as shown in Figure 3.7. The size and location of the square boxes were adjusted by
using the flow cytometry results obtained from pure samples of green (large) and red (small)
particles. Our aim was to isolate the green particles from the mixture of green and red particles
by pushing the larger particles in the upward direction. The green particles were separated
through the collection outlet with a 99.1% purity, and the red particles were separated through
the waste outlet with a 99.9% recovery rate.
18
Figure 3.7. Particle separation results for 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1. Flow cytometry graphs showing
the ratio of particles of different sizes collected at the collection (a) and waste (b), outlets respectively.
3.2.2. Characterization of inlets and outlets flow rate
In addition to the 𝜅-factor, we examined the effect of the flow ratio at the inlets (𝑄1, 𝑄2)
and outlets (𝑄3, 𝑄4). We first fixed the inlet flow ratio (𝑄1/𝑄2 = 1/9) so as to focus the
particles in the lower streamlines while varying the flow rate ratio at the outlets. We started
with an outlet flow ratio of 𝑄3/𝑄4 = 9/1. Under these flow conditions, all of the green (large)
particles were separated through the first outlet; however, a number of red (small) particles also
made their way into the green particle stream, resulting in only a 21.3% separation efficiency.
This low separation efficiency was attributed to the streaming flow produced by the TSAWs,
which disturbed the laminar flow streamlines inside the microchannel and resulted in mixed
particle sample collection at the outlets. The ratio of the sheath flow and sample flow heights
above the IDT was adjusted to achieve a high separation efficiency. We decreased the height
of the sheath flow and increased the height of the sample mixture flow after the IDT by
increasing the flow rate 𝑄4 and decreasing the flow rate 𝑄3 until the point(𝑄3/𝑄4 = 0.67/1).
Beyond this point, the green (large) particles passed through the second outlet along with the
red particles, reducing the recovery rate. As shown in Figure 3.8(a), the separation efficiency
increased as 𝑄3 decreased compared to 𝑄4. At 𝑄3/𝑄4 = 1. /1, the efficiency reached 90% and
exceeded 99% at 𝑄3/𝑄4 = 0.67/1.
We studied the effect of the inlet flow ratio by fixing the outlet flow ratio. Disruptions to
the flow separation by the streaming flow were avoided by selecting an outlet flow ratio at
19
which the separation reached a maximum (𝑄3/𝑄4 = 0.67/1). The flow rate at the inlet was
then varied to determine the effect of vertical particle focusing. As the inlets flow ratio 𝑄1/𝑄2
was increased from 1: 9 to 1: 1, as shown in Figure 3.8(b), the separation efficiency decreased
because the sample mixture flow height increased and the sheath flow decreased. The particles
were then dispersed under the sheath flow (not tightly focused). If the particles were dispersed
prior to reaching the separation zone, a non-uniform ARF effect on the particles would have
been observed.
Figure 3.8. Particle separation purity measures at different ratios of the inlet and outlet flow rates. (a) Percentage
of purity at various outlet flow ratios 𝑄3: 𝑄4 was varied from 9:1 to 0.67:1, holding the inlet flow ratio fixed
at 𝑄1: 𝑄2 = 1: 9. (b) Percentage of purities across inlet flow ratios of 𝑄1: 𝑄2 = 1: 9 to 1: 1, holding the outlet flow
ratio fixed at 𝑄3: 𝑄4 = 0.67: 1. (c), (d) Hemocytometer images obtained showing the particle size compositions
collected at different outlet and inlet flow ratios.
The flow ratios were adjusted accordingly to achieve a high separation efficiency. The
particle samples collected from the outlets at different inlet and outlet flow rate conditions were
quantitatively evaluated using the ImageJ software and a hemocytometer. Photographic images
showing a top view of the hemocytometer are presented in Figure 3.8(c, d) for different flow
ratios at the inlets and outlets.
The experimental results obtained at 𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1 are interpreted
as shown in the schematic diagram in Figure 3.9(a). As discussed earlier, the height of the
sheath flow 𝐻2 before the separation zone should be sufficiently greater than 𝐻1(𝐻1: 𝐻2 = 1: 9)
to pinch the green and red particles in the lower streamlines. By contrast, after the separation
zone, the height of the sheath flow streamlines 𝐻3 leaving through the first outlet should be
smaller than the height of the streamlines 𝐻4 exiting through the second outlet (𝐻3: 𝐻4 = : 3)
to avoid the effect of acoustic streaming on the separation efficiency.
20
Green particles with 𝜅 > 1 were affected by the ARF, whereas the red particles with 𝜅 <
1 were affected by the acoustic streaming. An experiment was performed to verify this
prediction. Particles were held stationary inside the microchannel. When the power was turned
on, the green particles with 𝜅 > 1 were pushed in the lateral and perpendicular directions due
to the horizontal and vertical components of the ARF;24 however, red particles with 𝜅 < 1 were
under the influence of the acoustic streaming flow,24 so they travlled along the circular
streamlines, displaying mixing behavior inside the microchannel, as shown in Figure 3.9(b).
Figure 3.9. (a) A side view schematic diagram illustrating the height of the sheath flow and the mixed particle
flow within sections of the microchannel. Particles with 𝜅 > 1 were affected by the ARF, whereas particles with
𝜅 < 1 were predominantly influenced by the streaming flow. (b) Top view of the experimental images, illustrating
the effects of SAWs on particles of different sizes.
3.2.3. Device operation at high flow rates
In addition to investigating the particles separation using the principal component of ARF
by uniquely focusing the particles in the vertical direction; we tested our device at higher flow
rates for the separation of 4.8 µm and 3.2 µm particles. We fixed the inlet flow ratio (𝑄1/𝑄2 =
1/9) and passively regulated the outlet flow ratio (𝑄3/𝑄4) by differential fluidic resistance,
and thus to reduce the number of external equipment. Flow rate at the first outlet 𝑄3 should be
greater than the flow rate at the second outlet 𝑄4, as we already characterized by giving a
negative pressure at one of the two outlets to avoid the effect of eckart streaming on the
separation efficiency. Figure 3.10 shows the purity and recovery measures at the first and
second outlets, respectively. Starting with the net flow rate, QNET = 5,000 µL/hr (cross sectional
velocity: V = 69.4 mm/s), we increased the QNET by doubling the prior one. As we increased
21
the QNET from 5,000 µL/hr to 80,000 µL/hr (V = 1110.4 mm/s) the purity gradually decreases
from 97.7% to 71.2%. In contrast, the recovery remains > 99% even at a much higher flow rate
(QNET = 80,000 µL/hr). Hemocytometer images of the sample collection at first and second
outlet are presented in Figure 3.11 and Figure 3.12 respectively for different net flow rates.
The gradual decrease in the purity with the increase of QNET was investigated. The effect of
streaming flow explained in Figure 3.9 might not play a major role in the decrease of purity at
higher flow rates; therefore we simulated the streamlines inside the microchannel to see the
behavior of streamlines at higher flow rates by varying the port size (inlets, outlets) and net
flow rate (QNET) .
Figure 3.10. Purity and recovery measures at different net flow rates QNET by holding the conditions (𝑄1/𝑄2 =
1/9) and (𝑄3 < 𝑄4) for the isolation of green, 4.8 µm and red, 3.2 µm particles.
Streamlines were simulated inside 250 µm wide microchannel with 1 mm inlets and
outlets port size, holding the inlets and outlets flow ratio constant (𝑄1/𝑄2 = 1/9 ; 𝑄3/𝑄4 =
0.67/1) as characterized before. As the 𝑄NET increased from 500µL/hr to 40,000µL/hr, there
was a formation of vortices at the first outlet due to the sudden change in the geometry.
Although 𝑄3 < 𝑄4 a few sample flow streamlines were still going with the sheath flow
streamlines resulting in a decrease of purity at the first outlet (see Figure 3.13). This effect of
vortices formation can be minimized by decreasing the port size. Simulation of streamlines
inside the microchannel with 500 µm diameter ports shows that there is a decrease of vortices
formation at the 1st outlet (See Figure 3.14). This problem can further be minimized by using
the hole size equal to the width of the microchannel in order to achieve high purity at both
outlets.
22
Figure 3.11. Hemocytometer images of sample collection at first outlet. Varied the net flowrates (𝑄NET) by fixing
the inlets and outlets flow ratio, 𝑄1 : 𝑄2 = 1 : 9 and 𝑄3 < 𝑄4 for the separation of green, 4.8 µm and red, 3.2 µm
particles.
Figure 3.12. Hemocytometer images of sample collection at second outlet. Varied the net flowrates (𝑄NET) by
fixing the inlets and outlets flow ratio, 𝑄1 : 𝑄2 = 1 : 9 and 𝑄3 < 𝑄4 for the separation of green, 4.8 µm and red,
3.2 µm particles.
23
Figure 3.13. Simulation of the streamlines inside 250 µm wide microchannel with 1mm punched hole at 𝑄1/𝑄2 =
1/9 and 𝑄3/𝑄4 = 0.67/1 for different net flowrates.
Figure 3.14. Simulation of the streamlines inside 250 µm wide microchannel with 500 µm punched hole at
𝑄1/𝑄2 = 1/9 and 𝑄3/𝑄4 = 0.67/1 for different net flowrates.
24
3.2.4. Tape based microchannel for the separation of particles
It should be noted that a simple, straight, manually fabricated tape-based microchannel
was used to successfully separate particles using the horizontal and vertical components of the
ARF. A mixture of 4.8 µm (green) and 2.0 µm (red) particles, at the characterized flow rate
conditions (𝑄1/𝑄2 = 1/9 ; 𝑄3/𝑄4 = 0.67/1) , was used to demonstrate the separation
process. Figure 3.15(a) shows the separation of particles inside the microchannel, supporting
with the experimental images of side view of the outlet pipes (see Figure 3.15(b)).
Figure 3.15. (a) A top view and (b) side view of the schematic with experimental images of the manually
fabricated scotch tape based microchannel.
3.3. Conclusions
We demonstrated the use of vertical hydrodynamic focusing to separate microparticles
in a straight microchannel using a single sheath flow. This technique is dissimilar from
previously reported methods of particle focusing using a pair of sheath flows that horizontally
focused particles in the center of a microchannel. The vertical focusing mechanism employed
here was implemented using a smaller device footprint compared to that used in horizontal
focusing techniques. Unlike previous acoustofluidic particles separation techniques, we used
the vertical component of the ARF to separate particles in a straight single layered
microchannel. The SAWs that leaked into water exerted a force along the vertical direction
(Fv) that far exceeded the force exerted in the horizontal direction (Fh), i.e. Fv ≅ 2.5 Fh. After
hydrodynamically pinching the particles in the downward streamlines (vertical direction), the
horizontal component of the ARF was used to slow down the motions of the larger particles
while at the same time pushing them upward using the vertical (major) component of the ARF.
We took advantage of the major component of the ARF to continuously separate 4.8 μm (green)
from 2.0 μm (red) and 3.2 μm (red) polystyrene microparticles. This method enables the highly
25
efficient (> 99% purity and recovery rates) separation of particles over a wide range of flow
rates using tens of milliwatts of power. Moreover, the device operates at higher flow rates (upto
80,000 µL/hr) with reasonable separation performance (purity 71.2%, recovery 99.4%).
Furthermore, we fabricated a straight microchannel using a strip of tape manually cut into the
desired shape. This method costs less than $1 for microchannel fabrication. This microchannel
fabrication procedure can replace standard MEMS and soft-lithography device fabrication
processes that cost more than $500 due to the use of expensive glass/chrome-based
photomasks. These characteristics render this particle separation technique a useful addition to
micro total analysis systems and point-of-care device.
26
Chapter 4. Acousto-chromatography based batchwise
concentration and separation of particles
4.1. Working Mechanism
A schematic of the pumpless microfluidic device shows a PDMS microchannel placed
on the SAW device which consists of a piezoelectric substrate (LiNbO3) with two IDTs
deposited on top of it (see Figure 4.1). A straight microfluidic channel with one inlet and one
outlet port was loosely aligned directly on top of the IDTs without plasma bonding. Firstly, a
few microliter of ethanol was injected through the inlet well with the help of pipette to make
the microchannel hydrophilic. Secondly, an inlet well was filled with the solution of different
diameter particles (red, green and blue fluorescent). The particles were flowed through the
microchannel due to the hydrodynamic pressure generated due to the height of the inlet well
and capillary action inside the microchannel. During SAW off, the particles mixture were
flowed through the microchannel and collected through the outlet which results in no separation
(see Figure 4.1(a)). Our study group and other researchers showed the interaction of TSAW
with water results in leaky SAW which radiate at an angle of 22o (with the microchannel side
wall based on Snell’s law) exerts force in the vertical direction (Fv) that far exceeded the force
in the horizontal direction (Fh), i.e. Fv ≅ 2.5Fh. Destgeer et al. illustrated the dependence of
ARF on particle diameter and TSAW in terms of a dimensionless number; 𝜅-factor, defined as
𝜅=𝜋𝑑 /𝑐 , where 𝑑 is the particle diameter, is the TSAW frequency, and 𝑐 is the speed of
sound in the fluid. This factor could be used to portray the particle motion under a TSAW. The
TSAW frequency of the IDT and the particle diameters were chosen such that the 𝜅>1 for red
particles and 𝜅<1 for green particles against first IDT. However, 𝜅>1 for green particles against
second IDT. The value of 𝜅-factor for smallest (blue) particles was less than one against both
the IDTs to ensure the least effect of ARF. During SAW on, two different sized particles with
𝜅>1 (red and green respectively) were trapped and concentrated beside first and second IDT
due to the vertical force component Fv generated from each IDT, while the smallest particles
with 𝜅<1 remained nearly unaffected by the ARF and filtered through the chromatography of
two different sized particles. Note that the horizontal (X-direction) component of the ARF Fh
was almost balanced with the drag force Fd on the particle; however, major (Y-direction)
component of the ARF Fv pushed the particles in the upward direction and seized its motion
27
through the microchannel based on the diameter of a particle and frequency of the particular
IDT. As a result, two different sized particles (red and green) were concentrated in specific
regions inside a single layered microchannel while the smallest (blue) particles were ultimately
collected through the outlet (see Figure 4.1(b)).
Figure 4.1. A schematic diagram showing the concentration and separation of three different diameter particles.
The device is composed of two parallel placed straight interdigitated transducer (IDT) patterned on the lithium
niobate (LiNbO3) substrate, a SiO2 layer and a straight PDMS microchannel loosely positioned on top without
plasma bonding. The top and side views of the device are showed as the concentration and separation zone is
enlarged during SAW off (a) and on (b) respectively. Particles slowed down by the horizontal component of ARF
Fh and concentrated by the vertical force component Fv depending on particle diameters.
4.2. Results and discussion
4.2.1. Concentration of particles by TSAWs
A sample solution containing similar sized particles was concentrated by using a single
IDT placed under a straight single-layered microchannel consisting of one inlet and one outlet.
A sample solution of 9.9 µm (green) diameter particles was pumped through the inlet with the
flowrate of 1000 µL/hr (Q1). When the device was turned off, all the green particles flowed
through the single-layered microchannel and collected through the outlet. It results in no
concentration of particles inside the microfluidic channel. The 𝜅-factor value for 9.9 µm
particles (green) was calculated to be 1.5. Therefore, when the device was actuated with AC
signal of 73 MHz frequency, the particles were pushed upward due to the vertical component
of the ARF and concentrated beside the IDT (without the assistance of microfabricated
membrane in a single-layered microchannel). However, the solution containing no particle was
h
h
LiNbO3
LiNbO3
PDMS
PDMS
IDT IDT
IDT IDT
Inlet well
Inlet well
PDMS
PDMS
LiNbO3
LiNbO3 Side view
Side view
Enlarged view
(a) SAW off
(b) SAW on
Top view
Top view Enlarged view
1st
IDT
> 1 < 1
> 1 < 1
2nd
IDT
28
collected at the outlet with flow rate of Q2. The hemocytometer images of inlet sample and
outlet sample can be seen in Figure 4.2.
Figure 4.2. A schematic diagram supported by a photographic image of particle concentration inside the
microchannel due to the combination of horizontal and vertical force components. Hemocytometer images of inlet
sample solution and outlet collected sample illustrated the efficient concentration of microspheres inside the
microfluidic channel.
In addition, the concentration of the polystyrene particles were characterized (w.r.t time) inside
the PDMS microchannel by changing the concentration of particles sample solution. With the
constant input power and flow rate as the sample concentration decreased the time required to
achieve the fixed concentration increases.
Figure 4.3. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 400 particles/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 1 sec to 30 sec. (c) Bar chart illustrate the decrease in the number
of particles per microliter in the sample solution results in the increase in required time to achieve fixed
concentration.
Outlet collected sample
Inlet Outlet
Inlet sample solutionTop view
Power on
SAWFrequency
73 MHz
Q1 Q2
Q1 (µL/hr)1000
0
10
20
30
40
50
0 10 20 30Time (s)
Inte
nsi
ty(a
.u)
0
5
10
15
400 200 100 20 10 1
110
115
120
Number of particles/µL
Tim
e (
s)
(b)
(c)
400 particles/µLInlet OutletTop view(a)
1 S
3 S
9 S
27 S
30 S
21 S
15 S
750 µm
29
We started with the sample concentration of 400 particles/µL. As the time increased from 1 sec
to 30 sec the intensity of the particles increased, which correspond to the increase in
concentration of particles inside the microchannel (see Figure 4.3(a,b)). Similarly, the
experiments were performed with the sample concentration of 200 particles/µL, 100
particles/µL, 20 particles/µL, 10 particles/µL and 1 particle/µL as shown in Figure 4.4 to
Figure 4.8 respectively. As the number of particles per microliter decreased from 400 to 1, the
time required to achieve the fixed intensity (concentration) increased from 0.38 min to 118 min
(see Figure 4.3(c)).
Figure 4.4. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 200 particles/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 6 sec to 60 sec.
Figure 4.5. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 100 particles/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 15 sec to 120 sec.
0
10
20
30
40
50
0 20 40 60Time (s)
Inte
nsi
ty(a
.u)
(b)
200 particles/µL
(a)6 S
12 S
18 S
24 S
30 S
36 S
42 S
48 S
54 S
60 S
0
10
20
30
40
50
0 20 40 60 80 100 120Time (s)
Inte
nsi
ty(a
.u)
100 particles/µL
(b)15 S
30 S
45 S
60 S
75 S
90 S
105 S
120 S
(a)
30
Figure 4.6. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 20 particles/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 1 min to 10 min.
Figure 4.7. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 10 particles/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 3min to 30 min.
(a)1 m
2 m
3 m
4 m
5 m
6 m
7 m
8 m
9 m
10 m
Time (m)
Inte
nsi
ty(a
.u)
(b)
200 particles/µL
0
10
20
30
40
50
0 5 10
20 particles/µL
(a)3 m
6 m
9 m
12 m
15 m
18 m
21 m
24 m
27 m
30 m
Time (m)
Inte
nsi
ty(a
.u)
(b)
200 particles/µL
0
10
20
30
40
50
60
70
0 10 20 30
10 particles/µL
31
Figure 4.8. Characterization of particle concentration inside the microchannel. The concentration of the sample
solution is 1 particle/µL. (a) Experimental images of top view of the microchannel and graph (b) showing the
linear increase in the particle concentration from 10 min to 120 min.
4.2.2. Particles trapping from extremely low concentrated solution
We demonstrated a highly sensitive SAW device for the capturing of particles from
extremely low concentrated particles solution. A device consisting of an inlet and outlet
mounted on the IDT, which was actuated by providing an AC signal of 73 MHz frequency. A
sample solution containing 10 particles/mL (9.9 µm, green) was injected through the inlet hole.
The experiment was run for two hours with the flow rate of 5 ml/hr; so, the target was to trap
100 particles. The images captured with time lap of 10 min revealed that almost all the particles
were successfully trapped and concentrated beside an IDT due to the principal component of
ARF, which seized its motion through the microchannel because of 𝜅>1 (see Figure 4.9). After
1 min of the process, 1 particle was captured. The number of captured particles was increased
to nearly half of the target after 1 hour. It kept on increasing with the passage of time and after
completion of 2 hours, 99 particles were captured out of 100 targeted particles. The particles
were counted manually and images are presented in Figure 4.9 with constant time gap after
increasing the size of the particles in ImageJ software for the sake of better visualization.
(a)10 m
20 m
30 m
40 m
50 m
70 m
80 m
90 m
100 m
110 m
60 m 120 m 0
5
10
15
20
25
30
35
40
0 20 40 60 80 100 120Time (m)
Inte
nsi
ty(a
.u)
(b)
1 particle/µL
32
Figure 4.9. Efficient particles trapping by the vertical component of ARF from extremely low concentrated
particles solution. The concentration of sample solution is 10 particles/ml. 99 out of 100 targeted particles were
captured in 2 hour.
4.2.3. A pumpless device for the concentration and separation of particles
In addition to characterization of particles trapping and concentration, we proposed a
pump-free acoustofluidic device for the concentration and separation of different sized particles
solution, which can be operated without a microscope. A mixture of microspheres was
concentrated and separated using two parallel placed IDTs beneath the straight microfluidic
channel: contained an inlet well and an outlet created by slicing the PDMS microchannel at the
opposite side of an inlet (see Figure 4.10). Experimental images showed the extent of particles
accumulation and isolation under the applied vertical force component of TSAW. After an
injection of 20 µL ethanol, a 100 µL solution contained three different sized particles (red, 12
µm; green, 4.8 µm & blue 2.1 µm) were filled in the inlet well with the help of pipette. The
particles flowed through the microchannel at a certain velocity which depends on the height of
the inlet well. When the device was off, the particles mixture flowed through the microchannel,
which results in no concentration or separation of particles. When the device was actuated 12
µm (red) particles were concentrated beside first IDT (F = 73 MHz) due to the vertical
component of ARF as 𝜅>1 for red particles and 𝜅<1 for green and blue particles. Similarly,
4.8 µm (green) particles were concentrated alongside second IDT (F = 140 MHz) due to the
vertical component of ARF as 𝜅>1 for green particles and 𝜅<1 for blue particles. The results
demonstrate that red and green particles were concentrated beside first and second IDT
respectively and the blue (smallest) particles leave through the microchannel. Figure 4.10(a)
shows the separation of blue, 2.1 µm particles through the chromatography of red, 12 µm and
1 min
10 min 20 min 30 min 40 min
50 min 60 min 70 min 80 min
90 min 100 min 110 min 120 min
Top view
OutletInlet
Frequency = 73 MHz
Power = 0.3 W
Flow rate = 5 mL/hrNumber of trapped particles = 1
= 13 = 18 = 29 = 37
= 45 = 52 = 61 = 68
= 72 = 82 = 89 = 99
Particle size = 9.9 µm
Sample concentration
= 10 particles / mL
33
green, 4.8 µm particles. After the process of 100 µL sample, a same quantity of ethanol was
filled with the pipette to wash the inlet well followed by the microchannel. Figure 4.10(b)
shows the concentration of red, 12 µm and green, 4.8 µm and particles at specific locations
inside the microchannel.
Figure 4.10. Photographic images of the particle concentration and separation experiment based on the stronger
force component of an applied ARF imposed by TSAWs. (a) Power on: blue particles passed through the
chromatography of red and green particles. (b) Power on: after washing with ethanol red and green particles
concentration is displayed. (c) Naked eye view of the concentrated (different diameter) particles inside the
microchannel. (d) Hemocytometer images obtained displaying the particle size compositions of concentrated and
separated particles.
4.2.4. Naked eye view of device operation, sample collection and separation
analysis
The proposed device can be functioned properly without a microscope and thus to reduce
the number of external equipment. The acousto-chromatography of different sized particles can
be seen with the naked eye similar to the pregnancy test kit in which two colored lines can be
seen on the piece of paper. As shown in Figure 4.10(c) red (12 µm) and green (4.8 µm) particles
were concentrated inside the single-layered microchannel producing a purified sample of 2.1
µm (blue) particles which was collected manually by placing a micro-pipe at the outlet as
shown in the photographic image of experimental setup (see Figure 4.11). The collection of
concentrated red and green particles was done after washing the microfluidic channel with the
ethanol as described earlier. Firstly, second IDT of 140 MHz frequency was turned off and the
green particles colony was collected at the outlet in the similar fashion as blue particles were
collected during the process. Secondly, first IDT of 73 MHz frequency was turned off and the
red particles colony was collected at the outlet. As a result, red, green and blue particles were
(A) Power on (During process)
(B) Power on (After process)
(D)
12 µm 4.8 µm 2.1 µm
(C)
750 µm
140 MHz73 MHz 70% 95% 100%
34
collected at the outlet for separation analysis. A hemocytometer was used to measure the purity
of each collected sample (see Figure 4.10(d)). Blue particles were collected with 100% purity,
followed by the collection of green and red particles with 95% and 70% purity respectively.
The reduction in the purity of concentrated samples, green and red particles can be attributed
to the secondary acoustic radiation force field formed due to the vibration of microspheres
which trapped the undesired particles resulting in the decreased of purity.
Figure 4.11. Photographic image of the experimental setup for naked eye view of the particles chromatography
and sample collection from the outlet for separation analysis.
4.3. Conclusions
We demonstrated a novel acoustofluidic technique for the concentration and separation
of different diameter particles inside a single layered microchannel without the assistance of
microfabricated membrane. Particles from an extremely low concentrated solution can be
detected, trapped and concentrated in highly efficient manner inside the microchannel. This
characteristic is particularly useful for the concentration and isolation of specific analyte such
as circulating tumor cells (CTCs) which are present in very small number (1-10/ml) inside the
blood of an infected person. The major component of ARF generated by TSAWs acted as a
wall for the concentration of selective particles inside the microchannel while allowing other
Collection outlet
outlet
Experimental setup Inletwell
AC signal
AC signal sliced
35
particles to filter through it. The proposed method is non-contact, noninvasive and label free.
These features are particularly significant for biochemical assays. The present acoustofluidic
chip does not require the induction of external pumping to inject the sample solution inside the
microchannel. Moreover, the operation of this device (trapping, concentration, separation) can
be visualized without a microscope. These characteristics of the device result in a reduction of
external tools and could be useful for point-of-care testing applications based on biological
sample preparation.
36
Bibliography
(1) Bhagat, A. A. S.; Bow, H.; Hou, H. W.; Tan, S. J.; Han, J.; Lim, C. T. Med. Biol. Eng.
Comput. 2010, 48 (10), 999–1014.
(2) Wyatt Shields IV, C.; Reyes, C. D.; López, G. P. Lab Chip 2015, 15 (5), 1230–1249.
(3) Yu, M.; Stott, S.; Toner, M.; Maheswaran, S.; Haber, D. A. Journal of Cell Biology.
2011, pp 373–382.
(4) Li, P.; Mao, Z.; Peng, Z.; Zhou, L.; Chen, Y.; Huang, P.; Truica, C. I.; Drabick, J. J.;
El-Deiry, W. S.; Dao, M.; Suresh, S.; Huang, T. J. Proc. Natl. Acad. Sci. 2015, 112
(16), 4970–4975.
(5) Gascoyne, P.; Satayavivad, J.; Ruchirawat, M. Acta Trop. 2004, 89 (3), 357–369.
(6) Gascoyne, P.; Mahidol, C.; Ruchirawat, M.; Satayavivad, J.; Watcharasit, P.; Becker,
F. F. Lab Chip 2002, 2 (2), 70–75.
(7) Nam, J.; Huang, H.; Lim, H.; Lim, C.; Shin, S. Anal. Chem. 2013, 85 (15), 7316–7323.
(8) Hewitt, Z.; Forsyth, N. R.; Waterfall, M.; Wojtacha, D.; Thomson, A. J.; Mcwhir, J.
Cloning Stem Cells 2006, 8 (3).
(9) Chen, Q.; Wu, J.; Zhuang, Q.; Lin, X.; Zhang, J.; Lin, J.-M. Sci. Rep. 2013, 3.
(10) Bryder, D.; Rossi, D. J.; Weissman, I. L. Am. J. Pathol. 2006, 169, 338–346.
(11) Abdallah, B.; Kassem, M. Gene Ther. 2008, 15, 109–116.
(12) Doi, D.; Samata, B.; Katsukawa, M.; Kikuchi, T.; Morizane, A.; Ono, Y.; Sekiguchi,
K.; Nakagawa, M.; Parmar, M.; Takahashi, J. Stem Cell Reports 2014, 2 (3), 337–350.
(13) Wang, S.; Wang, B.; Pan, N.; Fu, L.; Wang, C.; Song, G.; An, J.; Liu, Z.; Zhu, W.;
Guan, Y.; Xu, Z. Q.; Chan, P.; Chen, Z.; Zhang, Y. A. Sci Rep 2015, 5, 9232.
(14) Bhatia, S. N.; Ingber, D. E. Nat. Biotechnol. 2014, 32 (8), 760–772.
(15) Amini, H.; Lee, W.; Di Carlo, D. Lab Chip 2014, 14 (15), 2739–2761.
(16) Wu, Z.; Hjort, K. Micro Nanosyst. 2009, 1 (3), 181–192.
(17) Pamme, N.; Eijkel, J. C. T. T.; Manz, A. J. Magn. Magn. Mater. 2006, 307 (2), 237–
244.
(18) Kralj, J. G.; Lis, M. T. W.; Schmidt, M. A.; Jensen, K. F. Anal. Chem. 2006, 78 (14),
5019–5025.
(19) Choi, S.; Park, J.-K. Lab Chip 2005, 5 (10), 1161–1167.
(20) Jung, J. H.; Lee, K. H. K. S.; Lee, K. H. K. S.; Ha, B. H.; Oh, Y. S.; Sung, H. J.
Microfluid. Nanofluidics 2014, 16 (4), 635–644.
37
(21) Gascoyne, P. R. C.; Vykoukal, J. Electrophoresis. 2002, pp 1973–1983.
(22) Destgeer, G.; Lee, K. H.; Jung, J. H.; Alazzam, A.; Sung, H. J. Lab Chip 2013, 13 (21),
4210.
(23) Destgeer, G.; Ha, B. H.; Park, J.; Jung, J. H.; Alazzam, A.; Sung, H. J. Anal. Chem.
2015, 87 (9), 4627–4632.
(24) Destgeer, G.; Ha, B. H.; Jung, J. H.; Sung, H. J. Lab Chip 2014, 14 (24), 4665–4672.
(25) Chen, Y.; Wu, M.; Ren, L.; Liu, J.; Whitley, P. H.; Wang, L.; Huang, T. J. Lab Chip
2014, 16 (16), 3385–3616.
(26) Destgeer, G.; Cho, H.; Ha, B. H.; Jung, J. H.; Park, J.; Sung, H. J. Lab Chip 2016, 16
(4), 660–667.
(27) Destgeer, G.; Sung, H. J. Lab Chip 2015, 15 (13), 2722–2738.
(28) Shi, J.; Mao, X.; Ahmed, D.; Colletti, A.; Huang, T. J. Lab Chip 2008, 8 (2), 221–223.
(29) Shi, J.; Huang, H.; Stratton, Z.; Huang, Y.; Huang, T. J. Lab Chip 2009, 9 (23), 3354.
(30) Chen, Y.; Ding, X.; Steven Lin, S.-C.; Yang, S.; Huang, P.-H.; Nama, N.; Zhao, Y.;
Nawaz, A. A.; Guo, F.; Wang, W.; Gu, Y.; Mallouk, T. E.; Huang, T. J.; Ahsan
Nawaz, A.; Guo, F.; Wang, W.; Gu, Y.; Mallouk, T. E.; Jun Huang, T. ACS Nano
2013, 7 (4), 3306–3314.
(31) Destgeer, G.; Jung, J. H.; Park, J.; Ahmed, H.; Sung, H. J. Anal. Chem. 2017, 89 (1),
736–744.
(32) Destgeer, G.; Cho, H.; Ha, B. H.; Jung, J. H.; Park, J.; Sung, H. J. Lab Chip 2015, 16
(4), 660–667.
(33) Shilton, R. J.; Travagliati, M.; Beltram, F.; Cecchini, M. Adv. Mater. 2014, 26 (29),
4941–4946.
(34) Destgeer, G.; Jung, J. H.; Park, J.; Ahmed, H.; Park, K.; Ahmad, R.; Sung, H. J. RSC
Adv. 2017, 7 (36), 22524–22530.
(35) Schmid, L.; Weitz, D. A.; Franke, T. Lab Chip 2014, 14 (19), 3710–3718.
(36) Schmid, L.; Franke, T. Appl. Phys. Lett. 2014, 104 (13), 133501.
(37) Skowronek, V.; Rambach, R. W.; Schmid, L.; Haase, K.; Franke, T. Anal. Chem.
2013, 85 (20), 9955–9959.
(38) Collins, D. J.; Neild, A.; Ai, Y. Lab Chip 2015, 16 (3), 471–479.
(39) Ma, Z.; Collins, D. J.; Ai, Y. Anal. Chem. 2016, 88 (10), 5316–5323.
(40) Chen, Y.; Nawaz, A. A.; Zhao, Y.; Huang, P.-H.; McCoy, J. P.; Levine, S. J.; Wang,
L.; Huang, T. J. Lab Chip 2014, 14 (5), 916–923.
(41) Jung, J. H.; Destgeer, G.; Ha, B.; Park, J.; Sung, H. J. Lab Chip 2016, 77, 977–1026.
(42) Collins, D. J.; Alan, T.; Neild, A. Lab Chip 2014, 14 (9), 1595–1603.
38
(43) Collins, D. J.; Alan, T.; Neild, A. Appl. Phys. Lett. 2014, 105 (3), 33509.
(44) Shi, J.; Yazdi, S.; Lin, S.-C. S.; Ding, X.; Chiang, I.-K.; Sharp, K.; Huang, T. J. Lab
Chip 2011, 11 (14), 2319–2324.
(45) Nam, J.; Lee, Y.; Shin, S. Microfluid. Nanofluidics 2011, 11 (3), 317–326.
(46) Ding, X.; Peng, Z.; Lin, S.-C. S.; Geri, M.; Li, S.; Li, P.; Chen, Y.; Dao, M.; Suresh,
S.; Huang, T. J. Proc. Natl. Acad. Sci. 2014, 111 (36), 12992–12997.
(47) Guldiken, R.; Jo, M. C.; Gallant, N. D.; Demirci, U.; Zhe, J. Sensors 2012, 12 (12),
905–922.
(48) Hawkes, J. J.; Barber, R. W.; Emerson, D. R.; Coakley, W. T. Lab Chip 2004, 4 (5),
446–452.
(49) Chen, Y.; Wu, M.; Ren, L.; Liu, J.; Whitley, P. H.; Wang, L.; Huang, T. J. Lab Chip
2016, 16 (18), 3466–3472.
(50) Collins, D. J.; Ma, Z.; Han, J.; Ai, Y. Lab Chip 2017, 17 (1), 91–103.
(51) Witte, C.; Reboud, J.; Wilson, R.; Cooper, J. M.; Neale, S. L. Lab Chip 2014, 14 (21),
4277–4283.
(52) Kang, H. W.; Sung, H. J.; Lee, T.-M.; Kim, D.-S.; Kim, C.-J. J. Micromechanics
Microengineering 2009, 19 (1), 15025.
(53) Ha, B.; Park, J.; Destgeer, G.; Jung, J. H.; Sung, H. J. Anal. Chem. 2016, 88 (8), 4205–
4210.
(54) Rhee, M.; Valencia, P. M.; Rodriguez, M. I.; Langer, R.; Farokhzad, O. C.; Karnik, R.
Adv. Mater. 2011, 23 (12), 79–83.
(55) Shastri, A.; Mcgregor, L. M.; Liu, Y.; Harris, V.; Nan, H.; Mujica, M.; Vasquez, Y.;
Bhattacharya, A.; Ma, Y.; Aizenberg, M.; Kuksenok, O.; Balazs, A. C.; Aizenberg, J.;
He, X. 2015.