automated vitrification of mammalian embryos on a digital … › bitstream › 1807 › 65597 ›...
TRANSCRIPT
Automated Vitrification of Mammalian Embryos on aDigital Microfluidic Platform
by
Derek Geoffrey Pyne
A thesis submitted in conformity with the requirementsfor the degree of Master of Applied Science
Graduate Department of Mechanical and Industrial EngineeringUniversity of Toronto
c© Copyright 2014 by Derek Geoffrey Pyne
Abstract
Automated Vitrification of Mammalian Embryos on a Digital Microfluidic Platform
Derek Geoffrey Pyne
Master of Applied Science
Graduate Department of Mechanical and Industrial Engineering
University of Toronto
2014
This thesis presents the development of a digital microfluidic system to achieve auto-
mated sample preparation for the vitrification of mammalian embryos for clinical in
vitro fertilization (IVF) applications. This platform included micro devices fabrication,
an imaging system, a high voltage control system, and a LabVIEW interface. Individual
micro droplets manipulated on the digital microfluidic device were used as micro-vessels
to transport a single embryo through a complete vitrification procedure. The device
showed cell survival and development rates of 77% and 90%, respectively, which are
comparable to the control groups that were manually processed. Technical advantages
of this approach, compared to manual operation and channel-based microfluidic vitrifi-
cation, include automated operation, cryoprotectant concentration gradient generation,
and feasibility of loading and retrieval of embryos.
ii
Acknowledgements
I would like to thank my advisers, Professor Yu Sun and Professor Mohamed Abdelgawad,
for without their guidance and encouragement this thesis would not have been possible.
I wish to express my sincere appreciation and thanks to all the members of the Advanced
Micro and Nanosystems Laboratory at the University of Toronto. I would also like to
thank my collaborators Jun Liu (University of Toronto) and Waleed Salman (Assiut
University).
Finally, I would like to thank my parents for their continuing support, advice and
being my biggest fan.
iii
Contents
1 Introduction 1
1.1 Cryopreservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Digital Microfluidics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Merging Digital Microfluidics and Vitrification . . . . . . . . . . . . . . . 5
1.4 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Dissertation Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 System Setup 10
2.1 Device Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1.1 Photoresist Coating . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.1.2 UV Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.3 Developing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.4 Chromium Etching . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.1.5 Photoresist Removal . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.6 Dielectric Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.1.7 Teflon Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1.8 Assembled Devices . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2 Electrical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Control and Software System . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4 Imaging System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
iv
3 On-Chip Embryo Vitrification 26
3.1 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.2 Device Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 Embryo Loading and Retrieval . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Cryoprotectant Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.5 Thawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.6 Measurement Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.7 Vitrification Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.8 Volume Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.9 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4 Partially Filled Electrodes 47
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Modelling and Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.3 Experimental Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5 Summary 57
5.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.2 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
A Detailed Fabrication Recipe 60
B Selected LabVIEW Code 66
Bibliography 66
v
List of Tables
3.1 Summary of vitrification results. . . . . . . . . . . . . . . . . . . . . . . . 35
4.1 Parameters used in numerical simulations. . . . . . . . . . . . . . . . . . 49
vi
List of Figures
1.1 Vitrification process involves sequentially bathing the embryo/oocyte in
cryoprotectant baths according to a strict timing protocol. Cryoprotec-
tant concentration increases in later baths and the embryo is plunged
immediately after exiting the final bath. The first bath is typically called
Equilibrium Solution (ES), and later baths called Vitrification Solution
(VS). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Overview of fundamental digital microfluidic operations. Droplets can be
transported around the device for mixing or movement to another seg-
ment of the device. Smaller droplets can be dispensed from larger holding
reservoirs. Droplets of larger volume can be split into two daughter droplets. 4
1.3 Droplet mixing is achieved by first merging droplets by actuating on to
a common electrode. The merged droplet is then transported around the
device until mixing is homogeneous. Red dye was used to help visualize
this process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Schematic showing differences between manual vitrification approach, which
requires manual pipetting between mediums in cryoprotectant stage, and
the digital microfluidic (DMF) approach, which moves the embryo be-
tween mediums on chip. The chip automates the high skill portion of the
procedure providing labor cost savings and opportunities for parallelization. 7
2.1 Overview of system elements and central LabVIEW interface. . . . . . . 10
vii
2.2 Overview of device fabrication. Photolithography was used to pattern
chromium electrodes, chemical vapour deposition was used to deposit pary-
lene C, and Teflon wass finally spin coated on the device. . . . . . . . . . 11
2.3 Photoresist is spin coated on the device at 3000 rpm for 30s. . . . . . . . 12
2.4 Photoresist is patterned by exposing to UV light through a photomask for
10s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.5 During development areas of photoresist exposed to UV light are washed
away in the developer. The desired pattern is then left in the photoresist
layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.6 During etching, areas of the chromium layer not protected by photoresist,
are dissolved in the etchant. This transfers the design pattern from the
photoresist layer to the chromium layer. . . . . . . . . . . . . . . . . . . 15
2.7 Photoresist is removed using a stripper solution in an ultrasonic bath. The
ultrasonic bath helps removes photoresist from all sections of the device. 16
2.8 Parylene C is used as a dielectric layer and deposited using a chemical
vapour deposition process with a dedicated coater. . . . . . . . . . . . . . 17
2.9 Teflon is applied by spin coating a uniform layer, and then baking to
remove solvents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.10 Assembled device is held in place using a set of 3D printed parts. Top
ITO slide is connected to ground using an alligator clip. . . . . . . . . . . 18
2.11 Schematic of electrical system showing relay array control and high voltage
generation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.12 Relay array containing 24 mechanical relays, connection to LabVIEW sys-
tem, and connection to ribbon cables to device. . . . . . . . . . . . . . . 20
2.13 Electrodes are interfaced to relay system using an edgeboard connector.
Allows for fast changing between devices. Parylene is removed from elec-
trodes to ensure good connection. . . . . . . . . . . . . . . . . . . . . . . 21
viii
2.14 Overview of LabVIEW programming environment. The Front Panel is the
user interface and holds controls and indicators usable by the operator.
The Wiring Diagram controls data flow and is done using a graphical
programming paradigm where data flows through wires instead of local
variables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.15 Overview of droplet control LabVIEW elements. A pattern of buttons
is used for manual control in the shape of the present device. Droplet
sequence control is accomplished by reading electrode sequences from an
excel file. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.16 Overview of imaging apparatus. Zoom and focus are motorized allowing
them to be controlled from within LabVIEW. . . . . . . . . . . . . . . . 24
2.17 LabVIEW front panel highlighting imaging components. Zoom and focus
are controlled either by sliders on the right side of the screen, or by a toggle
allowing quick switching between predefined low and high zoom settings.
Videos are recorded with a toggle and files saved using timestamps as their
name. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.1 Chip design showing regions for vitrification medium dispensing and em-
bryo inlet/outlet. The top ITO slide is placed on the device in a manner
that exposes portions of the top electrodes in the dispensing reservoir and
the leg of the T-shape to allow for medium and embryo loading respectively. 27
3.2 Embryo is input and extracted by actuating electrodes at edge of top glass
slide. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.3 Overview of general approach for mixing on a digital microfluidic platform. 30
3.4 Schematic showing implementation of mixing protocol using this device
design. Daughter droplet containing embryo in step 4 is identified manually. 31
ix
3.5 Mixing profile showing generation of ES medium and VS medium. Exper-
imental droplet concentrations were found by using image processing to
measure the droplet volumes before and after each mixing step. . . . . . 32
3.6 (a) Embryo (red circle) contained in culture medium (CM) droplet. (b)
Embryo droplet mixed with VS droplet. (c) Droplet split into two droplets
(left contains embryo). (d) Droplet containing embryo is kept and other
droplet is sent to waste. Process is repeated to increase VS concentration. 33
3.7 Sample healthy and failure cases for survival rate based on morphology
before and after freezing. . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.8 Sample healthy and failure cases for development rate based on culturing
for an additional 24-48 hours after freezing. . . . . . . . . . . . . . . . . . 35
3.9 Embryo cell volume measurements for a typical (a) human and (b) mouse
protocol on chip. Volumes were normalized to initial volume. The initial
volume dip in the human protocol matches the volume dip over the mouse
protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.10 Comparison of mouse and human vitrification protocols. [38, 47–52] . . . 38
3.11 Implementation of common vitrification protocols on a digital microfluidic
chip with a single dispensing reservoir. Timings and concentrations are
shown in (a), and the generalized mixing curve is shown in (b). . . . . . . 39
3.12 A storage ring could be used to store multiple embryos each in their own
droplet. This would allow embryos to be loaded all at once, and then
individually processed when needed. . . . . . . . . . . . . . . . . . . . . . 42
3.13 Schematic showing droplet transfer to removable freezing device. (a) Em-
bryo initially inside device, (b) droplet moved outside of closed structure,
(c) droplet moved using open digital microfluidic electrodes, (d) droplet
transported onto freezing device, and (e) freezing device removed and di-
rectly plunged into liquid nitrogen. . . . . . . . . . . . . . . . . . . . . . 44
x
3.14 (a) Droplet is in contact with substrate in air and (b) floating over a small
oil layer when in an oil bath. . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.1 (a) Schematic of partially filled electrodes providing space for additional
on-chip tools or as a window for imaging. (b) Example designs of partially
filled electrode configurations considered. Electrodes are 1mm x 1mm (c)
Image of droplet on series of partially filled electrodes. . . . . . . . . . . 48
4.2 Simulation results. (a) Droplet half space mesh and swept uniform mesh
on droplet surfaces. (b) Actuation force on droplet for conventional elec-
trode on leading and trailing droplet surfaces. Reverse actuation force is
generated on trailing droplet surface as backward interface begins to move
onto electrode. (c) Induced forces increase linearly with electrode fill ra-
tio, which was changed by varying the width of the horizontal bars in the
electrode. (d) Force is independent of vertical location of removed area
from electrode. The leading edge of the droplet is fixed at the midpoint of
the electrode for panels (c) and (d). . . . . . . . . . . . . . . . . . . . . . 51
4.3 Simulation results: electrode design with crescent-like filled areas at the
entrance and exit of the electrode produces increased force at beginning
and end of droplet motion to create initial force to ensure droplet motion
is generated. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
xi
4.4 Experimental comparison of electrode designs. In experiments, maximum
droplet actuation frequency was measured at different fill percentages for
normal and improved designs. Number of horizontal bars in the electrode
was kept constant, thus reducing the bar width, reduces electrode fill per-
centage. Please note that reduction in maximum actuation frequency is
almost proportional to the reduction in the bar width similar to the force
reduction simulations. Experiments were conducted with deionized water
at 75 V rms and 15 kHz by actuating droplet back and forth across a series
of 5 electrodes at increasing speed until droplet motion could not keep up
with actuation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.5 (a) Single mouse embryo morphology on partially filled Cr electrodes. Up-
per image uses bright field transmission DIC imaging showing superior
detail compared to reflection microscopy imaging used in lower image.
(b) RBC viability measured with different osmolarity by counting cells
through unfilled regions on chip. N = 150 600 for each osmolarity point. 56
B.1 Camera communication is opened using IMAQ commands. Some initial
options are written such as packet size, gamma control, and exposure time.
This camera feed is fed into the image capturing loop. . . . . . . . . . . . 67
B.2 To simplify camera adjustments, color balancing is done automatially us-
ing a single button on the front panel. When the ’Auto Balance’ is pressed
the gain and white balance are automatically adjusted using functions on
the camera. These settings can be rebalanced during experimentation if
lighting or sample conditions change. . . . . . . . . . . . . . . . . . . . . 68
xii
B.3 Inside the main imaging loop, frames are taken from the camera and dis-
played on an imaging window in the front panel. When the ’Record’ button
is initially pressed, a AVI container file using an MPEG compressor is cre-
ated using the current timestamp as the filename to ensure that there are
no filename conflicts. After this initially loop, if the ’Record’ button is
still pressed, individually frames are added to the AVI container file. A
timer is also used to display the frame rate on the front panel. . . . . . . 69
B.4 Communication with the signal generator is achieved over USB communi-
cation. Standard VISA commands are used to set the function, voltage,
and frequency values. Communication is sent to the signal generator when-
ever either the voltage, or frequency control values on the front panel are
changed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
B.5 Macro control is initiated by reading in excel files containing the list of
droplet actuations for each command. These files are named to match their
function and each button on the front panel corresponds to an individual
excel file. Once the file is opened it is fed into the main droplet control loop. 71
B.6 In the main droplet control loop, references to each electrode button are
first build into an array so that they are indexed. Each line of excel file
is then read and the corresponding electrodes turned on. Options are
available to leave the reservoir electrodes on at all times to ensure that the
large reservoir droplet is not moved. The time between actuations can also
be set using a front panel control. Communication with an LCR meter
was also initiated to sense droplet locations after each step. However, the
droplet motion was found to be very robust, making the extra complexity
of droplet sensing not necessary. . . . . . . . . . . . . . . . . . . . . . . . 72
xiii
Chapter 1
Introduction
1.1 Cryopreservation
Cryopreservation is a key technology in biology and clinical practice. It is a process
where substances, cells, or even whole tissue are cooled to low enough temperatures
that all enzymatic and chemical activity is essential stopped, alowing the preservation
of the sample for an indefinite amount of time. The first successful pregnancy following
cryopreservation was reported in 1983 [1]. Stem cells [2], sperms [3, 4], and embryos [5]
are now routinely frozen and preserved for use at a later time. Patients who undergo
therapeutic procedures that can place their fertility at risk, such as chemotherapy, have
the option of preserving their reproductive cells, such as sperms or oocytes, for future
use through in vitro fertilization techniques (IVF) [6–9]. Furthermore, extra fertilized
embryos after an IVF procedure can also be frozen for use at a later time. The length
of time frozen has been shown not to have a significant impact on clinical pregnancy,
miscarriage, implantation, or live birth [10, 11]. However, in these cases the number of
viable cells for preservation can be low making the survivability rate and reproducibility
of cryopreservation techniques critical. In particular, preservation of embryos or oocytes
is challenging since these cells are highly sensitive, and the cell number is very small.
1
Chapter 1. Introduction 2
Fig 1: Vitrification schematic
ES VS1 VS2 Freeze
Figure 1.1: Vitrification process involves sequentially bathing the embryo/oocyte in cry-oprotectant baths according to a strict timing protocol. Cryoprotectant concentrationincreases in later baths and the embryo is plunged immediately after exiting the finalbath. The first bath is typically called Equilibrium Solution (ES), and later baths calledVitrification Solution (VS).
The two commonly used cryopreservation techniques for freezing embryos are the
slow freezing method and the vitrification method. Both techniques aim to minimize the
damage caused by freezing that is largely due to the formation of intracellular ice crystals
that can produce a mechanical shear and rupture the cells [12]. Conventionally, cells are
frozen through the slow freezing method where cells are placed in a large freezer that
can accurately control the freezing rate down to liquid nitrogen temperatures, with low
concentrations of cryoprotectants [13]. During slow freezing extracellular water freezes
away from the embryo, using a seeding technique, which creates an osmotic gradient that
draws water out of the cell until it finally freezes without the formation of intracellular ice
crystals [14]. This procedure requires sophisticated equipment to control the freezing rate,
which ranges between 0.3 and 1.0◦C/min, and produces a relatively poor survivability
rate [15, 16].
On the other hand, vitrification offers an alternative approach in which cells are frozen
at extremely high rates, usually by directly plunging the sample into liquid nitrogen, af-
ter bathing them in a sequence of high concentration cryoprotectants [17]. Vitrification
reduces intracellular ice formation, which is the primary cause of cell death, by freez-
ing the sample in a glass-like state before the molecules have a chance to form crystal
structures. This results in a higher cell survival rate after thawing compared to conven-
Chapter 1. Introduction 3
tional slow freezing without the need for a seeding procedure or a programmable freezer
[15, 18]. However, vitrification requires precise washing sequences and timings in each
cryoprotectant medium since higher concentrations are used and there is a significant
risk of toxicity if overexposed (Fig. 1.1). The process is expensive in terms of technical
skills required. In IVF clinics, processing an embryo/oocyte in cryoprotectant medium
typically costs a highly skilled embryologist 10 to 15 minutes.
Three factors are essential when implanting and designing a vitrification protocol;
cooling rate, viscosity of mediums used, and volume of medium frozen [14, 19]. The
cooling rate needs to be as high as possible to ensure that vitrification occurs before
crystal formation. The viscosity of the medium also works to enable vitrification as
with higher concentrations of cryoprotectants the glass transition temperature is raised,
allowing vitrification to happen earlier in the cooling process. The volume of the medium
surrounding the embryo while freezing is also critical as smaller volumes allow faster heat
transfer and thus a higher chance of vitrification. Surrounding containers and the thermal
conductivity of the holding devices can also work to slow the cooling/warming rate.
1.2 Digital Microfluidics
Traditionally, microfluidic research has largely centered around channel-based systems in
which fluids are controlled through microchannels using pumps, valves, and mechanical
mixers [20]. These continuous systems are well suited for applications such as diagnos-
tics for blood, urine and saliva [21]. Digital microfluidics, or electrowetting on dielectric
(EWOD), uses an alternative approach in which individual liquid droplets are manipu-
lated in discrete steps. This creates a new set of design opportunities better suited for
handling low sample number applications with a high degree of control. Larger cells, or
solids can also be handled without the risk of clogging channels [22]. An array of elec-
trodes is used to control droplet movement. Electrode configurations are not application
Chapter 1. Introduction 4
ON ON
ON ON ON ON ON
ON ON ON ON
Transport
Dispensing
Splitting
1 2 3
1 2 3
1 2 3
Figure 1.2: Overview of fundamental digital microfluidic operations. Droplets can betransported around the device for mixing or movement to another segment of the device.Smaller droplets can be dispensed from larger holding reservoirs. Droplets of largervolume can be split into two daughter droplets.
specific and thus can be generalized and reconfigured for different applications. Elec-
trode widths typically vary from 0.5-2.0 mm (or larger for reservoir droplets) creating
droplets in the nL to mL range [23]. These small liquid volumes require small reagent
consumptions, and allow for faster reaction rates. Loading and unloading of samples can
also be easily done by simply pipetting droplets on the edge of the device or by building
the device with a capillary tube inside the structure [24].
Droplet motion in a digital microfluidic system is typically created by applying a high
voltage across a dielectric coated gap in which the liquid droplet is contained. This creates
a large electric field within the insulating dielectric layer, which in turn produces a surface
charge on the leading edge of the droplet. This creates a net horizontal force that moves
the droplet towards the actuated electrode [23, 25, 26]. This force is only generated when
a section of the droplet overlaps the actuated electrode. This leads to the requirement
that the gap between adjacent electrodes be as small as possible to ensure droplet motion.
However, this has the added benefit that droplet force is generated locally on the device
allowing droplets to be isolating from another with no dependence on each others motion.
More details modelling the force generation in a digital microfluidic system are shown in
Chapter 1. Introduction 5
Chapter 3.
Similar to the evolution of the logic gate in the digital world, using the binary oper-
ation of a single electrode, basic operations can be built and performed to accomplish a
large amount of liquid handling tasks. Transport is the basic operation allowing droplet
transfer between sections or modules on the device. Merging of droplets is achieved by
actuating multiple droplets on the same electrode, and mixing is achieved by transport-
ing the droplet in a pattern until its composition is homogenous. Larger droplets can
be split into two daughter droplets by simultaneously stretching a droplet in opposite
directions, and dispensing is similarly performed by stretching a single side of a larger
reservoir droplet until a smaller daughter droplet is dispensed. These basic actions are
shown in Figure 1.2 and are the fundamental operations used to build more complicated
laboratory sample processing tasks. Figure 1.3 also shows a mixing operation using dyes
to visualize droplet mixing.
1.3 Merging Digital Microfluidics and Vitrification
Digital microfluidics is a power tool for sequential sample processing and has been used
in tasks such as PCR, cell culture, and immunoassays [27–29]. In this work, digital mi-
crofluidics is used, for the first time, to automate embryo preparation for the vitrification
procedure, aiming to lower the high labour cost and ultimately helping further spread
the use of vitrification in IVF clinics.
The key to automating the vitrification process is to replicate the washing and timing
steps of a given protocol while also keeping complete control of the embryo (as the
sample population may be quite small making each individual cell very critical). Digital
microfluidics is proposed to be uniquely positioned to address this task since droplets on
the digital microfluidic platform can act as micro-vessels to move an embryo and subject
it to a series of cyroprotectants of different concentrations, as required by IVF vitrification
Chapter 1. Introduction 6
Figure 1.3: Droplet mixing is achieved by first merging droplets by actuating on toa common electrode. The merged droplet is then transported around the device untilmixing is homogeneous. Red dye was used to help visualize this process.
Chapter 1. Introduction 7
Manual C
ryop
rote
ctan
t S
tage
DMF Chip
Ext
ract
ion
Sta
ge
Free
zing
S
tage
ES
VS VS
VS
embryo
micropipette
liquid nitrogen
Figure 1.4: Schematic showing differences between manual vitrification approach, whichrequires manual pipetting between mediums in cryoprotectant stage, and the digitalmicrofluidic (DMF) approach, which moves the embryo between mediums on chip. Thechip automates the high skill portion of the procedure providing labor cost savings andopportunities for parallelization.
Chapter 1. Introduction 8
protocols (Figure 1.4). Compared to channel-based microfluidic vitrification [30–32], the
device reported in this work does not undesirably park the embryo to a confined area
and has less intricacy of embryo introduction and retrieval onto and from the device.
By keeping the embryo in a single droplet, as opposed to microchannels or wells, the
system is also able to constantly track and control the embryos locations throughout the
procedure to avoid cell loss.
1.4 Motivation
Overall, by creating a standardized platform for embryo vitrification several majors ben-
efits are realized. Firstly, this platform would drastically lower the cost of fertility preser-
vation hopefully opening up this procedure to a wider range of patients. This is especially
relevant as with the development of new cancer treatments creating an increase in the
number of surviving patients, greater attention is being focused on providing a high
quality of life to the survivors [7]. Secondly, by eliminating the need for a highly skilled
embryologist, a higher number of clinics and labs could perform this procedure open-
ing the geographic availability. Finally, by automating the procedure, variability from
manual operation would be eliminated. This would standardize the implementation of
vitrification protocols allowing for stronger comparisons between timings and mediums.
If adopted, this would significantly aid researchers in progressing cryopreservation work.
1.5 Dissertation Outline
This thesis is organized into the following chapters. Chapter 2 describes the digital
microfluidic platform, including electronics and software, and the fabrication process used
to produce the devices. Chapter 3 presents a system for automation of the vitrification
process for mammalian embryos along with initial mouse embryo testing. Chapter 4
presents a study on partially filled electrode designs including finite element simulations,
Chapter 1. Introduction 9
preliminary experimental results, and suggestions for improved designs. Finally, Chapter
7 summarizes the impact of this system.
Chapter 2
System Setup
Building a digital microfluidic platform involves coordinating several different systems
and pieces of equipment together in order to generate precise, fast droplet controls. On
top of this basic function, systems also need to be incorporated to help automate other
elements of the users tasks to quicken and streamline operation, as certain experiments
are time sensitive. The microchips themselves also require cleanroom fabrication tech-
niques, which posses an additional challenge. Here we outline our platform, software
and fabrication techniques, which aim to address this issues and create a robust, efficient
experimentation environment. Figure 2.1 shows an overview of the platform including
Fig 1: Overview system
LabVIEW controller
digital microfluidic chip
high voltage signal generation
imaging system
Figure 2.1: Overview of system elements and central LabVIEW interface.
10
Chapter 2. System Setup 11
Figure 1.5: Chip fabrication overview
top glass plate
bottom glass plate
embryo
Teflon
parylene C
Cr electrode
ITO
liquid droplet
embryo
top glass plate
bottom glass plate
Teflon parylene C Cr electrode ITO
liquid droplet Figure 2.2: Overview of device fabrication. Photolithography was used to patternchromium electrodes, chemical vapour deposition was used to deposit parylene C, andTeflon wass finally spin coated on the device.
sample holder, electrical system, imaging system, and LabVIEW control interface.
2.1 Device Fabrication
Since digital microfluidic chips require a micro scale gap between adjacent electrodes to
create droplet motion, cleanroom fabrication techniques were used. A cleanroom is a
low particulate laboratory that reduces the amount of particulate that can settle on a
sample during fabrication and create defects in the device. Lithography is the major
cleanroom procedure in our chips and involves the transfer of a geometric pattern from
a photomask to a chromium layer through the use of photosensitive materials [33]. Our
chips contain an electrode gap of 20 µm, which is smaller then can be produced in printed
circuit board processes, yet is relatively large on lithography standards. The Emerging
Communications Technology Institute (ECTI) facilities at the University of Toronto were
used which is a class 1000 cleanroom (maximum 1000 particles ≤ 0.5 µm per cubic foot of
air). Glass slides pre coated with 100 nm of chromium were used as the starting substrate
(Deposition Research Labs Inc., St. Charles, MO). This eliminated the need for metal
evaporation of chromium saving both time and money. The lithography process, which
patterns the chromium layer, and the additional layer deposition processes, is outlined
Chapter 2. System Setup 12
below.
2.1.1 Photoresist Coating
Photoresist is the key component in the lithography process. It has the unique property
that after exposure to UV light, its solubility to a developer compound changes. This
means that after developing, either the exposed, or unexposed, areas will dissolve in
the developer depending on the type of photoresist used. Positive photoresist references
photoresist in which the exposed areas are washed away by the developer, and negative
photoresist references photoresist in which the unexposed areas are washed away by
the developer. This enables patterns to be transferred onto a photoresist coating by
selectively exposing different areas of the photoresist by using a photomask. Initially, the
substrate was primed with Hexamethyldisilazane (HMDS) before spin coating Shipley
S1811 photoresist (3000 rpm, 30s). The sample was then soft backed at 115◦C for 2
minutes to harden the layer by removing solvents. At this point a uniform thickness
layer of photoresist has been deposited on the chromium coated glass slides (Fig. 2.1.1).
Fig 2: Spin coater
bottom glass plate
chromium positive photoresist
(a) (b)
(a) Photoresist Coating
Fig 2: Spin coater
bottom glass plate
chromium positive photoresist
(a) (b)
(b) Spin Coater
Figure 2.3: Photoresist is spin coated on the device at 3000 rpm for 30s.
Chapter 2. System Setup 13
2.1.2 UV Exposure
The sample is next exposed to UV light for 10s through a photo mask printed on a
transparency at high resolution (Pacific Arts and Design, Markham ON). The areas of
photoresist that are exposed become more soluble to the developer and are thus removed
during developing. For our devices, the exact dimension of the gap is not critical, only
that is small enough to create reliable motion. If the gap is larger, motion can be slightly
slower, and may not be reliable in some cases. However, if the gap is too small, and there
is a connection between adjacent electrodes, a short circuit is formed. In this case these
two electrodes act as a single electrode and the device is not functional. Thus to avoid
this risk the sample is overexposed, making the gap larger then on the photomask, to
ensure no short circuits are formed (Fig. 2.1.2).Fig 3: Mask aligner
bottom glass plate
UV EXPOSURE
chromium positive photoresist
photomask
(a) (b)
(a) UV Exposure
Fig 3: Mask aligner
bottom glass plate
UV EXPOSURE
chromium positive photoresist
photomask
(a) (b)
(b) Mask Aligner
Figure 2.4: Photoresist is patterned by exposing to UV light through a photomask for10s.
2.1.3 Developing
To remove the exposed areas of photoresist, the sample is submerged in MF-321 developer
for approximately 2 minutes. The timing in this step is forgiving as sufficient time only
needs to be given to remove all the photoresist from the exposed areas, and the unexposed
areas will only slowly start to dissolve themselves. After some experience development
Chapter 2. System Setup 14
completion can be judged by eye. The sample is rinsed with deionized water to remove
any residual developer. The sample is then hard baked at 115 ◦C for 60 s to further harden
the photoresist by removing solvents. At this point the pattern has been transferred from
the photomask to the photoresist and is visible to the user (Fig. 2.1.3).
Fig 4: Developer
bottom glass plate
chromium positive photoresist
(a) (b)
(a) Developing
Fig 4: Developer
bottom glass plate
chromium positive photoresist
(a) (b)
(b) Photoresist Developer
Figure 2.5: During development areas of photoresist exposed to UV light are washedaway in the developer. The desired pattern is then left in the photoresist layer.
2.1.4 Chromium Etching
An acidic solution, CR-4 etchant, is used that is specifically designed to etch chromium
patterns. The sample is etched for approximately 3 minutes in CR-4 chromium etching
solution, with some gentle agitation to speed up the process. Similar to developing,
the timing in this process is not critical. Over-etching can result in undercut features
and wider gaps, which dont affect the devices. This etching removes chromium from
all areas not protected by photoresist, exposing the glass substrate underneath. The
sample is rinsed with deionized water to remove any residual etchant. The samples can
now be investigated under a microscope to ensure that the devices are defect free. If
short circuits or deformed electrodes are found, the devices should be disposed off and
the fabrication process restarted as there is no point in moving forward with defective
devices (Fig. 2.1.4).
Chapter 2. System Setup 15
Fig 5: Etching
bottom glass plate
chromium positive photoresist
(a) (b)
(a) Chromium Etching
Fig 5: Etching
bottom glass plate
chromium positive photoresist
(a) (b)
(b) Chromium Etchant
Figure 2.6: During etching, areas of the chromium layer not protected by photoresist,are dissolved in the etchant. This transfers the design pattern from the photoresist layerto the chromium layer.
2.1.5 Photoresist Removal
Finally, photoresist still remaining on top of the chromium pattern is removed with AZ-
300T stripper. Samples are placed in a beaker of stripper and placed in an ultrasonic
bath to ensure complete removal of photoresist. Samples are then well rinsed with deion-
ized water, as the stripper is viscous and can leave residue if not rinsed properly. The
lithography process is now complete and the desired chromium pattern produced. The
lithography process should be completed in a single session as the properties of the pho-
toresist can change if left for even a couple of days, due to solvent evaporation. However,
the samples can now be left in this state for as long as needed before continuing on to
the next coatings (Fig. 2.1.5).
2.1.6 Dielectric Coating
Parylene C is used as a dielectric coating to stop electrolysis. It is a conformal coating and
applied in a vapour deposition process at room temperature using a dedicated coating
machine. Parylene C is purchased as a solid in small pellets. A weighed amount of
Parylene is placed into the coater and the thickness of the resulting layer is dictated
Chapter 2. System Setup 16
Fig 6: Ultrasonic bath
bottom glass plate
chromium
(a) (b)
(a) Photoresist Removal
Fig 6: Ultrasonic bath
bottom glass plate
chromium
(a) (b)
(b) Ultrasonic Bath
Figure 2.7: Photoresist is removed using a stripper solution in an ultrasonic bath. Theultrasonic bath helps removes photoresist from all sections of the device.
by this weight, as the coater will automatically run until all the available Parylene is
deposited and cannot deposit a partial coating. A two stage heating process is used to
heat the Parylene into a monomer before it enters the chamber with the samples and is
deposited as a clear uniform layer. Since this deposition is at room temperature, there are
no concerns about melting of any components of the sample. This means that tape can
also be placed over the outer electrode connections to keep these areas free of Parylene.
Parylene in these regions would need to be removed anyway to insure a strong connection
with the edge board connector, so using tape during this coating speeds up the process.
Overall, a 2 µm thick layer of dielectric is produced (Fig. 2.1.6).
2.1.7 Teflon Coating
Teflon AF1600 is used as a final hydrophobic coating and applied by spin coating at 1600
rpm for 60s and then baking at 160 ◦C for 10 minutes. The Teflon mixture is a liquid
composed of 1 % w/w Teflon AF1600 in FC-40. Teflon AF1600 is purchased as a solid
and dissolved in the liquid FC-40 by baking on a hotplate at 80 ◦C for several hours
beforehand. The top ground slide of the devices is a glass slide pre coated with indium
tin oxide (ITO). ITO is used, as it is a clear conductor and a robust coating. These slides
Chapter 2. System Setup 17
Fig 7: Parylene C coater
bottom glass plate
chromium parylene C
(a) (b)
(a) Parylene C Coating
Fig 7: Parylene C coater
bottom glass plate
chromium parylene C
(a) (b)
(b) Vapour Deposition Coater
Figure 2.8: Parylene C is used as a dielectric layer and deposited using a chemical vapourdeposition process with a dedicated coater.
are also coated in Teflon using the same technique (Fig. 2.1.7).Fig 8: Teflon spin coating
bottom glass plate
chromium parylene C Teflon
(a) (b) top glass plate
(a) Telon Coating
Fig 8: Teflon spin coating
bottom glass plate
chromium parylene C Teflon
(a) (b) top glass plate
(b) Teflon Spin Coating
Figure 2.9: Teflon is applied by spin coating a uniform layer, and then baking to removesolvents.
2.1.8 Assembled Devices
Finally, the devices are assembled by placing the top ITO slide, which acts as a ground
plane, over the patterned device using two pieces of double sided tape as a spacer
(Fig. 2.10). There is a balance to achieve strong droplet motion while also being able
to dispense and split droplets. A large gap can produce stronger motion but make dis-
Chapter 2. System Setup 18
Fig 9: Assembled device
ground connection
digital microfluidic device
double-sided tape spacers
sample holder
Figure 2.10: Assembled device is held in place using a set of 3D printed parts. Top ITOslide is connected to ground using an alligator clip.
pensing very difficult, as the droplet surface is harder to split. The ITO slide is offset
so that one end is hanging and can be connected to ground using an alligator clip. The
other edge of the slide is positioned so that the outermost device electrode is half covered
and half exposed. This allows droplets to be pipetted on the edge of the ITO slide and
actuated into the device either to fill reservoirs, or input samples.
2.2 Electrical System
The electrical system is the core component in a digital microfluidic system as it drives
the functionality of the system; compared to other components built around control and
monitoring of the device. Droplet actuation is created by creating a large electric field
across the device. This causes charge to build up in the dielectric layer underneath the
contact line of the droplet. A corresponding surface charge is then formed in the droplet
creating a net electrostatic force in the horizontal direction. Very little current flows
through the device, by design, yet a large voltage is needed to generate the necessary
electric field. The signal is created using a function generator (Agilent 33522A) and a high
voltage amplifier able to provide up to 100 Vrms at kHz frequencies (Trek PZD350A).
The driving signal is typically 15 kHz with an amplitude of 50-80 Vrms. Care needs to be
Chapter 2. System Setup 19
Fig 10: Electrical system
function generator
voltage amplifier
relay array edgeboard connector
device electrodes
LabVIEW controller
Figure 2.11: Schematic of electrical system showing relay array control and high voltagegeneration.
made when dealing with high voltage as there is the potential for significant harm. The
voltage amplifier current limit should be set low to minimize this risk, especially since
high current is never needed in the functioning device. Lowering the current limit also
reduces the damage caused by electrolysis in a malfunctioning device or electrode.
It is essential that the voltage is computer controlled to allow for reliable, fast droplet
motion. An array of mechanical relays was used for this task. Two state relays were used
connecting each electrode to either ground or high voltage. In their off state the relays
were connected to ground. Since these relays required more then 10 V and significant
current (approximately 50 mA per relay when on) a transistor was used to connect
the relay control switch to a DC power supply. The transistors were controlled with a
digital signal provided by the LabVIEW DAQ. An overview of this system is shown in
Figure 2.11.
A total of 24 relays were used which provided ample electrodes for this papers ap-
plications (Fig. 2.12). Also, in certain chip designs multiple electrodes can be connected
together to conserve relays if either parallel motion is desired, or each area of the device
is sufficiently distanced to not interfere with each other. Electrodes were connected to
the relays by an edge board connector and a ribbon cable (Fig. 2.13). This allowed for
chips to be quickly connected by simply sliding them into the connector. The connector
Chapter 2. System Setup 20Fig 11: Relay circuit
connection to ribbon cables
relay circuit boards (24 relays total)
NI DAQ connector
Figure 2.12: Relay array containing 24 mechanical relays, connection to LabVIEW sys-tem, and connection to ribbon cables to device.
had a pitch of 2.54 mm allowing for 27 electrodes to be placed on each edge of the chip.
This allowed for 2-3 devices to fit on a single glass slide depending on the number of
electrodes used.
2.3 Control and Software System
LabVIEW, the graphical programming environment, was chosen as the programming
language for our platform. LabVIEW excels in interfacing with laboratory equipment,
quickly creating a graphical user interface, and integrating with National Instruments
array of physical tools (such as analog/digital input and output devices). For digital
microfluidics, this fit exactly the needs of the platform and allowed more time to be
spent on device design and experimentation as apposed to a control system built in a
more traditional environment, such as C++, where the specifics of device control and
the user interface would have to be dealt with.
Programming in LabVIEW is done through two windows; the front panel and the
Chapter 2. System Setup 21
Fig 12: Edgeboard connector
(a) (b)
open connected (a) Open
Fig 12: Edgeboard connector
(a) (b)
open connected (b) Closed
Figure 2.13: Electrodes are interfaced to relay system using an edgeboard connector.Allows for fast changing between devices. Parylene is removed from electrodes to ensuregood connection.
wiring diagram as shown in Figure 2.14. The front panel is the user interface and is
where all buttons, controls, and imaging windows are placed. Elements are added to
this window from a tools palette and are resizable and movable on the front panel.
The wiring diagram is where all logic and programming is done. All components on
the front panel have a corresponding block in the wiring diagram. Data flows through
the wiring diagram by connecting these blocks together with wires, instead of the more
conventional approach of using named variables. Data is manipulated through function
blocks that have both inputs and outputs for wires. Loops or case structures are created
by surrounding sections of code in a rectangular fence. Wires can also run into and out
of these structures. This style of programming allows for additional tools to be quickly
added to the front panel and to be ran in parallel in the wiring diagram. This allows the
system to quickly shrink and expand without spending time building a scaling interface
and infrastructure.
Droplet motion is controlled using a National Instruments Data Acquisition (DAQ)
board, which is connected to the relay array. The LabVIEW control interface is high-
lighted in Figure 2.15. Control of each electrode is done through 24 buttons on the front
Chapter 2. System Setup 22Fig 13: Front panel, wiring
(a) (b)
front panel wiring diagram (a) Front Panel
Fig 13: Front panel, wiring
(a) (b)
front panel wiring diagram (b) Wiring Diagram
Figure 2.14: Overview of LabVIEW programming environment. The Front Panel is theuser interface and holds controls and indicators usable by the operator. The WiringDiagram controls data flow and is done using a graphical programming paradigm wheredata flows through wires instead of local variables.
panel. This allows the user to manually actuate selected electrodes and also move, scale,
and colour the buttons to better represent the layout of the current device. Although
these buttons are controls, their state can also be programmatically controlled within the
wiring diagram. This allows the user to also design automated sequences of droplet actu-
ations. This system reads in excel files that contain an ordered list of droplet actuations
(more than one electrode can be actuated at once) and executes them with a fixed time
interval, which can also be adjusted on the front panel. The idea is to break down droplet
protocols into a series of smaller operations such as mixing, dispensing, transport, and
rearranging. This allows the user to execute the complete protocol by simply executing
these smaller segments in order. This has allowed for quicker testing and debugging as
the user is able to assess the situation after each segment and make any small manual
adjustments needed, or even repeat a segment if necessary. Electrodes are left on after
each segment to prevent the droplets from floating away. Segments can also be stopped
if needed using the Stop button.
Control over the electrical signal is also performed by communication over USB with
Chapter 2. System Setup 23
Fig 14: Front droplet control
signal control
manual droplet control
droplet sequence
control timing
interval
Figure 2.15: Overview of droplet control LabVIEW elements. A pattern of buttons isused for manual control in the shape of the present device. Droplet sequence control isaccomplished by reading electrode sequences from an excel file.
the function generator. LabVIEW contains pre made functions for standard protocols
over USB, Ethernet, and GPIB. The user is able to turn the function generator on and
off, change the voltage level, and change the frequency. This allows for quick changes
to be made on the fly during experimentation as different liquids and geometries require
different voltage strengths to achieve motion while avoiding electrolysis.
2.4 Imaging System
Different imaging systems can be used in a digital microfluidic platform. Transmission
microscopy often allows for higher resolution imaging and even the use of fluorescent
imaging. However, if conventional metal electrodes are used, these block the imaging
path and prevent the imaging of anything in the droplet over an electrode. This can be
circumvented by fabricating transparent electrodes, usually out of ITO, although this is
not common practice. An alternative is to use reflection microscopy in which the light
source is normal to the electrodes; passing through the objective lens, reflecting off the
device, and passing back into the objective lens. This produces a strong metallic reflection
off of the electrodes allowing for imaging of its contents. In our platform reflection
Chapter 2. System Setup 24
Fig 15: Imaging overview
camera
motorized zoom/focus
2x objective lens
Figure 2.16: Overview of imaging apparatus. Zoom and focus are motorized allowingthem to be controlled from within LabVIEW.
microscopy was used with a motorized optical system allowing for computer controlled
adjustment of both the zoom and focus as shown in Figure 2.16. Since this papers
primary application dealt with mouse embryos, this allowed quick switching between
magnifications during experiments to either monitor droplet motion at low magnification
or embryo morphology at high magnification.
All of these controls were built into the LabVIEW platform as shown in Figure 2.17.
Video recording was also provided with each video saved using a timestamp as its file
name. This allowed the user to quickly record multiple videos during an experiment
and then organize and change the name of important videos after experimentation was
complete.
Chapter 2. System Setup 25
Fig 16: Labview imaging
imaging window
video recording
motorized zoom control
focus control
Figure 2.17: LabVIEW front panel highlighting imaging components. Zoom and focusare controlled either by sliders on the right side of the screen, or by a toggle allowingquick switching between predefined low and high zoom settings. Videos are recordedwith a toggle and files saved using timestamps as their name.
The sample was also held in place with a set of 3D printed parts. These parts
were made to work with a salvaged microscope stage. This allowed for devices to be
quickly secured to the stage using only two screws and to be manually moved during
experimentation in the horizontal plane for imaging different areas of the device.
Chapter 3
On-Chip Embryo Vitrification
3.1 Materials
Mouse embryos were gathered from the Canadian Mouse Mutant Repository (CMMR;
Toronto, ON). Embryos were produced by superovulating a female and were gathered
2.5 days past conception, which corresponds to most embryos being in the 8-cell stage.
Vitrification solution usually contains antifreezing agents or cryoprotectant, such as
dimethyl sulfoxide (DMSO), some small molecular size glycols (e.g., ethylene glycol), or
sucrose [34]. A mixture of several cryoprotectants is often used to reduce the individ-
ual specific toxicity as well as using both permeable and impermeable mediums [15]. A
combination of DMSO and sucrose was used to follow the protocol used by the embryo
suppliers. The vitrification solution (VS) was made by diluting DMSO in serum-free
KSOM medium (EMD Millipore, Billerica, US) at 33% concentration, with 1.0 M su-
crose. The equilibrium solution (ES) was at half concentration of VS (i.e., 16.5% DMSO
+ 0.5 M sucrose). VS was preloaded on the DMF chip before each experiment. The first
mixing step, which mixes the VS with embryo culture medium (i.e., serum-free KSOM),
generates the ES.
26
Chapter 3. On-Chip Embryo Vitrification 27
Fig 4: Overview chip design
waste#reservoir#
embryo#inlet/outlet#
dispensing#reservoir#
reservoir#inlet#
mixing/spliOng#
waste&reservoir&
embryo&inlet/outlet&
dispensing&reservoir&
reservoir&inlet&
mixing/spli6ng&
edge#of#top#ITO#slide#
edge#of#top#ITO#slide#
Figure 3.1: Chip design showing regions for vitrification medium dispensing and embryoinlet/outlet. The top ITO slide is placed on the device in a manner that exposes portionsof the top electrodes in the dispensing reservoir and the leg of the T-shape to allow formedium and embryo loading respectively.
3.2 Device Design
Voltages applied to actuate droplets were 55-75 Vrms at 15 kHz. Cyroprotectant droplets
were actuated inside silicone oil (2.0 cSt, Gelest Inc., Morrisville, PA) to reduce friction
and evaporation. Different regions of the device were designed to achieve the general
digital microfluidic fucntions of transporting, mixing, dispensing, and splitting droplets.
A large reservoir was used to hold and dispense the high concentration cryoprotectant
medium, as shown in Figure 3.1. This reservoir was split up into many sections to
handle variations in liquid volume in the reservoirs as droplets are dispensed during the
vitrification protocol. The second reservoir was used as a waste reservoir and, thus, was
split into two large electrodes only as less control was needed.
A central inverted T-shaped array of electrodes was used for droplet transport, mixing,
and splitting. Electrodes in this array were interdigitated to allow droplet overlap with
adjacent electrode and increase electrodynamic forces applied on droplets. Top electrode
in the leg of the T-shape was an input/output region where half of the edge electrode was
Chapter 3. On-Chip Embryo Vitrification 28
Fig 5: Droplet input/output
embr
yo
inpu
t em
bryo
re
triev
al
2 1 3
2 1 3
top
slid
e ed
ge
mic
ropi
pette
Figure 3.2: Embryo is input and extracted by actuating electrodes at edge of top glassslide.
exposed out of the ITO slide to enable embryo loading. For embryo loading, the embryo-
carrying droplet was pipetted on the exposed half of the electrode and then actuated
into the device through the covered half [35]. For extraction, the embryo-carrying droplet
was moved to this edge electrode where it bulges out of the device and is retrieved by
a standard micropipette. This same mechanism was used to fill the dispensing reservoir
before the embryo and reagents were loaded.
3.3 Embryo Loading and Retrieval
As shown in Figure 3.2, to input an embryo, a small embryo-containing droplet was
pipetted onto the loading electrode and then actuated into the device. This technique
minimized exposure of the embryo to outside air. Extraction of the embryo was completed
in the opposite manner by transporting the embryo-containing droplet to the edge of the
device and retrieved by a micropipette. Once the embryo was extracted from the device,
it was directly frozen in the micropipette inside liquid nitrogen.
Chapter 3. On-Chip Embryo Vitrification 29
3.4 Cryoprotectant Mixing
An embryo is input into the device in a small droplet of embryo culture medium, and
100% cryoprotectant is input into the device in larger volumes (reservoir inlet in Fig. 3.1).
The cryoprotectant bathing procedure is then performed through a serial mixing/splitting
process (Fig. 3.4 and Fig. 3.5). This is accomplished by mixing the embryo-containing
droplet with a vitrification solution droplet (VS), thus increasing the concentration of
cryoprotectant around the embryo. The resulting droplet is then split into two smaller
droplets with the daughter droplet containing the embryo identified and kept, while
the other droplet is moved to the waste reservoir. After the first mixing step the droplet
reaches 50% cryoprotectant concentration (i.e., equilibrium solution or ES), the embryo is
kept in the ES droplet for 10 minutes. Then the cryoprotectant concentration is increased
again by droplet mixing and splitting. Contrary to ES medium, embryo volume sharply
decreases in VS medium and does not recover (Fig. 3.9). The overall mixing profile
generated with a single dispensing reservoir is shown below in Eqn. 3.1.
C(n) = 100
(1− 1
2n
)%, (3.1)
where C(n) is the concentration of the droplet, and n is the number of mixing steps. This
protocol mimics a typical two-step human embryo/oocyte protocol. However, mouse
embryos are typically frozen with only a single step protocol where the embryo is directly
transferred to VS medium and then frozen. On our chip this corresponds to simply
removing the 10-minute waiting period in the ES medium step. Both protocols were
performed but the results presented were done with the mouse embryo timings to follow
precisely the protocol provided to us by CMMR.
After complete transfer of the embryo into the VS medium, the droplet containing
the embryo is moved toward the edge to be collected by a micropipette (Fig. 3.2), and
then plunged into liquid nitrogen. To verify success of the vitrification process using
Chapter 3. On-Chip Embryo Vitrification 30
Fig 2: Droplet mixing
1. Initial droplets
2. Droplets merged
3. Droplet mixed
4. Droplet Split
Figure 3.3: Overview of general approach for mixing on a digital microfluidic platform.
digital microfluidics, embryos vitrified on device were thawed back and confirmed to
have recovered in volume and have healthy morphology.
Contrary to conventional vitrification protocols and manual operation, which subject
embryos to sudden changes in medium concentration, the digital microfluidic approach
gradually increases the VS medium concentration, (Fig. 3.5), which is generally accepted
by IVF practitioners to be more benign to embryos due to lower osmotic stress [36]. This
gradual medium concentration increase is not feasible to achieve in manual operation.
3.5 Thawing
The thawing procedure for embryo vitrification is much simpler then the freezing process.
Embryos were thawed by plunging in a bath of culture medium with 1.0 M of sucrose.
The sucrose helps to draw the cryoprotectants out of the embryo by osmotic pressure
Chapter 3. On-Chip Embryo Vitrification 31
1. Embryo input using edge of top ITO slide
2. Droplet dispensed from reservoir
3. Droplets merged and mixed until homogeneous
4. Droplet split into two daughter droplets
5. Embryo containing droplet kept, remaining droplet moved to waste
Figure 3.4: Schematic showing implementation of mixing protocol using this device de-sign. Daughter droplet containing embryo in step 4 is identified manually.
Chapter 3. On-Chip Embryo Vitrification 32Fig 3: Mixing Curve
(b)
(d)
VS
(a)
(c)
CM
0102030405060708090
100
0 2 4
drop
let c
onc.
(%)
number of mixing steps
theoretical
measured
ES VS
(b)
(d)
VS
(a)
(c)
CM Figure 3.5: Mixing profile showing generation of ES medium and VS medium. Exper-imental droplet concentrations were found by using image processing to measure thedroplet volumes before and after each mixing step.
Chapter 3. On-Chip Embryo Vitrification 33
Fig 3: Mixing Curve
(b)
(d)
VS
(a)
(c)
CM
0102030405060708090
100
0 2 4
drop
let c
onc.
(%)
number of mixing steps
theoretical
measured
ES VS
(b)
(d)
VS
(a)
(c)
CM
Figure 3.6: (a) Embryo (red circle) contained in culture medium (CM) droplet. (b)Embryo droplet mixed with VS droplet. (c) Droplet split into two droplets (left containsembryo). (d) Droplet containing embryo is kept and other droplet is sent to waste.Process is repeated to increase VS concentration.
to minimize toxicity. The embryo is left in this bath for approximately 10 minutes over
which its volume slowly increases back to its original size. After this point the embryo
is transferred to culture medium without sucrose and is returned to the incubator. It
is generally advised that the embryo not be used for any additional applications until it
has had several hours to equilibrate in the incubator.
The thawing temperature gradient is equally important to the freezing gradient and
should be as quick as possible. The micropipette containing the embryo in liquid nitrogen
is thus transferred to the initial sucrose bath in one smooth, quick motion. The sucrose
bath is also of sufficient size to have a large enough thermal mass to quickly warm
the micropipette tip. The same thawing procedure was used for manual and digital
microfluidic operation. Since this procedure is much simpler and not the main focus of a
vitrification protocol, it was held constant to put more emphasis on the cryoprotectant
preparation protocol prior to freezing.
Chapter 3. On-Chip Embryo Vitrification 34
Fig 7: Embryo morphology
Healthy
16 cells
Sept 27 continued
Sept 30 - S: 0/2 (0%)
Oct 7 - S: 4/5 (80%), D: 3/5 (60%)
Oct 9 - S: 1/1 (100%), D: 1/1 (100%)
Sept 27 - S: 4/5 (80%), D: 2/4 (50%)Oct 11 - S: 2/2 (100%), D: 2/2 (100%)
Unhealthy
16 cells compacted
8 cells
shrunken
ruptured cells
darkened cells
Figure 3.7: Sample healthy and failure cases for survival rate based on morphology beforeand after freezing.
3.6 Measurement Scheme
Two measures were used to evaluate the performance of digital microfluidic vitrification,
including survival rate and development rate. Survivability was measured by examining
the morphology of the embryo before and after freezing [37] (Figure 3.7). Embryos
were considered unhealthy if they had an abnormal shape, membrane damage, leakage
of cellular content or degeneration of their cytoplasm [38]. The development rate was
determined by culturing survived embryos for an additional 24-48 hours after freezing
and thawing (Figure 3.8). If the cell number within the embryo increased or it developed
to the blastocyst stage, it was counted as developed. Control samples of non-vitrified
embryos were also cultured to identify the base development rate of the mouse embryo
population. Only embryos morphologically judged to be healthy were used for either
manual or digital microfluidic testing. Only embryos that had healthy morphology after
freezing were cultured following similar procedures to other vitrification studies [39].
Chapter 3. On-Chip Embryo Vitrification 35
Fig 8: Development embryo
Oct 9 - S: 1/1 (100%), D: 1/1 (100%)Oct 9 - S: 1/1 (100%), D: 1/1 (100%)
Sept 27 continuedSept 27 continued
Sept 27 - S: 4/5 (80%), D: 2/4 (50%)Sept 27 - S: 4/5 (80%), D: 2/4 (50%) Oct 7 continuedOct 7 continued
Healthy Unhealthy Before
Freezing After 1-2
Days Before
Freezing After 1-2
Days
developed to 16 cell compact
stage
developed to blastocyst
all cells dead
cells no longer growing and
coloring is dimmer
Figure 3.8: Sample healthy and failure cases for development rate based on culturing foran additional 24-48 hours after freezing.
3.7 Vitrification Results
Table 3.1 summarizes the results, showing comparable survival and development rates
between manual and automated trials. However, with a larger population size I hypoth-
esize the digital microfluidic chip to produce a higher survival rate due to its gradient
generation.
Table 3.1: Summary of vitrification results.Survival Rate Development Rate
Control (Non vitrified) 100% (14/14) 93% (13/14)Manual 73% (11/15) 91% (10/11)DMF Chip 77% (10/13) 90% (9/10)
3.8 Volume Monitoring
Additionally, since the embryo is constantly imaged on video, its volume can be measured
throughout the procedure and used to measure the quality of both the embryo and the
protocol. Figure 3.9 shows one such assessment.
Chapter 3. On-Chip Embryo Vitrification 36
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
0 2 4 6 8 10
volu
me
(nor
mal
ized
)
time (minutes)
liqui
d ni
troge
n fre
ezin
g
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
1
0 0.25 0.5 0.75 1 1.25
volu
me
(nor
mal
ized
)
time (minutes)
ES# VS# VS#50µm
liqui
d ni
troge
n fre
ezin
g
(a) (b)
Figure 3.9: Embryo cell volume measurements for a typical (a) human and (b) mouseprotocol on chip. Volumes were normalized to initial volume. The initial volume dip inthe human protocol matches the volume dip over the mouse protocol.
3.9 Discussion
Previous vitrification studies for embryos (4-16 cell stages) reported survival rates in the
range of 80-100% [5]. Vitrification using digital microfluidics in this work showed a sur-
vival rate of 77% that is similar to the manual vitrification trials (survival rate of 73%).
This lower survival rate, compared to the results in the literature, can be mainly at-
tributed to our use of a micropipette (vs. vitrification straw) inside liquid nitrogen. The
micropipette tip is a standard plastic pipette for embryo manipulation and had an inner
diameter of 125 µm (The STRIPPER micropipetter, Origio). Much research has gone
into developing different mechanical structures (e.g., straw-type carriers [40], cryotube
[41], Cryotop [42], and mesh-type carriers [43]) to increase the heat transfer rate [14].
Structures have also been developed and modeled with the specific measurement of suc-
cess being a high heat transfer coefficient [44]. In these designs materials are chosen with
a higher heat transfer coefficient and the liquid volume containing the embryo is mini-
mized. Some researchers have also directly dropped the embryo containing droplet into a
liquid nitrogen bath to maximize the freezing rate [45]. Using these devices would involve
transferring the embryo from the DMF chip manually onto the device (e.g., vitrification
straw) with minimal liquid volume. Since this would introduce an extra manipulation
Chapter 3. On-Chip Embryo Vitrification 37
step by hand and this work focuses on proving the feasibility of using digital microfluidics
for embryo processing, the embryos were directly frozen inside the micropipette tip in
this work.
Using micropipette tips in liquid nitrogen was not ideal and negatively affected the
survival rate; however, since it was held constant between manual and DMF trials, the
relative survival rates can still be compared. Additionally, the development rate, which
was measured using embryos that survived freezing, was high and comparable with other
vitrification studies. This measurement in some ways removes the effect of the limited
manual embryo handling skills and shows the potential of the microfluidic device for
automated processing of embryos for vitrification.
Due to the programmable characteristic, one key benefit to the digital microfluidic
approach is the ability to implement/test a number of vitrification protocols for efficacy
comparisons. Table 3.10 lists example human and mouse vitrification protocols, all using
different cryoprotectants, number of mediums, and timings. However, at their core, all
of these protocols involve the controlled increase in cryoprotectant concentration over
a given period of time with the initial equilibrium solution typically containing half the
cryoprotectant concentration as the full strength vitrification solution. This is convenient
for DMF device design as it always produces a 50% concentration after the first mixing.
This means that a typically two-step protocol involving an initial 50% concentration ES
step, following by a short 100% VS step, can be easily realized by following the mixing
curve shown in Figure 3.5 with a pause after the first mixing step to allow the embryo
to reach equilibrium.
Figure 3.9 shows how most two-step protocols follow the same mixing curve, only
differing by their timings. This shows that although the protocols in Table 3.10 involve
significantly different amounts of manual manipulation, when performed on the digital
microfluidic chip, they follow the same mixing procedure with the only difference lying
in their timings. For our current trials, a mouse protocol provided by CMMR was used
Chapter 3. On-Chip Embryo Vitrification 38
CMMR DMSO, sucrose 8-cell VS
Dhali et al. DMSO, EG, sucrose zygote to morulae VS
Khosravi-Farsani et al. EG, sucrose oocyte VSi
Irvine DMSO, EG, sucrose 2PN to blastocyst
Origio EG, PG, sucrose ii 4-cell to 16-cell VS
Kitazato DMSO, EG, sucroseii 2PN to 8-cell VS1iii VS2iii
iOnly step containing sucroseiiExact concentrations not giveniiiSame medium but different bath
ES3ES2ES1
15time (minutes)Protocol Cryoprotectants 1 3 5 7 9 11 13Stage
ES
ES
ES VSi
ES
mou
se&
human&
Figure 3.10: Comparison of mouse and human vitrification protocols. [38, 47–52]
which involved a single mixing step. This allowed us to better conduct our manual trials
as it required less manipulation; however, more complicated protocols can be readily
performed by adding more dispensing reservoirs on chip and filling them with lower
concentrations of cryoprotectant. This would allow for a high number of producible
concentrations, especially in the low concentration range. Multiple reservoirs could also
be used to implement protocols using different cryoprotectant compositions throughout
their procedure [46].
One limitation of the digital microfluidic platform was handling culture mediums
containing high serum concentrations. Serum contains a mixture of proteins that can
absorb on the Teflon coated surfaces of the device, eventually accumulating to the point
that the surface becomes hydrophilic, making droplets immovable [53]. Some strategies
have been developed to help overcome this problem, such as the use of Pluronic additives
[54], silicone oil baths [55], and superhydrophobic surfaces [56]. Pluronic additives were
avoided in our work as embryos are highly sensitive to additives. Superhydrophobic
surfaces were also avoided as they require significant additional fabrication efforts. A
silicone oil bath was used which did increase droplet movability; however, this approach
did not work with sufficient effectiveness for conventional embryo culturing mediums that
contain high serum concentrations. Therefore, in this work, a serum free culture medium
was chosen for this proof-of-principle study. Further work will focus on droplet movability
improvement for serum-containing culture medium and the automation of transferring
embryos onto vitrification devices (e.g., straw or Cryotop).
Chapter 3. On-Chip Embryo Vitrification 39
c(t)
t t1 t2 t3
100%
c1
Protocol c1 t1 t2 t3 t4
CMMR 100% 0.5 1 - -
Dhali et al. 50% 1 3 3.5 4
Irvine 50% 1 8 9 10
Origio 50% 1 10 10.5 11
Kitazato 50% 1 12 13 14
(a) (b)
t4
(a) Mixing curve c(t)
t t1 t2 t3
100%
c1
Protocol c1 t1 t2 t3 t4
CMMR 100% 0.5 1 - -
Dhali et al. 50% 1 3 3.5 4
Irvine 50% 1 8 9 10
Origio 50% 1 10 10.5 11
Kitazato 50% 1 12 13 14
(a) (b)
t4 (b) Protocol timings
Figure 3.11: Implementation of common vitrification protocols on a digital microfluidicchip with a single dispensing reservoir. Timings and concentrations are shown in (a),and the generalized mixing curve is shown in (b).
Chapter 3. On-Chip Embryo Vitrification 40
Additionally, some platforms have also incorporated temperature control during the
vitrification process to improve embryo culture conditions [46]. Due to the planar nature
of the digital microfluidic platform, and the relatively thin glass substrate used, a warming
plate could easily be integrated in the future. This is especially critical if the system is
expanded to handle multiple embryos at once as this may involve storing multiple embryos
on the chip initially, before vitrifying each embryo individually. This would require these
embryos to remain in storage on the device for longer periods of time in which case
maintaining optimal culture conditions would be critical.
Overall, a platform was designed and tested to automate the liquid handing and
timing portion of a vitrification procedure. The platform is general in its design allowing
clinicians and researchers to easily change protocols or develop new ones. This is of
particular importance as vitrification protocols are not presently standardized making
a platform that could do this highly valuable. Once a common platform is adopted,
it is easier for clinicians to develop protocols by comparing their findings with another
clinicians using the same platform. Embryo quality is also more easily assessed when
the same tools are used. Manual vitrification involves a large amount of manipulation
by pipette which leads to high variance between clinics. The manner that washing is
performed, the amount of liquid aspired with the embryo and the speed transferring
between baths all affect the outcome. In order for vitrification to be used widespread, a
system like this would be highly beneficial.
The programability of the platform also enables researchers to attempt more compli-
cated protocols that were not realistic when the embryo needed to be manually transferred
between baths. Typical procedures now consist of at most 4-5 steps, with this platform
procedures can be easily broken down into smaller segments and the complexity greatly
increased. In addition, continuous video monitoring could allow for closed-loop feedback
protocols. In other words, if image processing tools were added such that the embryo
volume is monitored in real-time in the LabVIEW interface, this information could be
Chapter 3. On-Chip Embryo Vitrification 41
used to adjust protocol timings on the fly. For instance, as the embryo volume shrinks,
this information could be used to forecast the time at which it will reach minimum vol-
ume and be removed from the device at this point. This modeling could be done without
modeling the physiology of the embryo but instead by fitting a curve to previous em-
bryo experiments and using the same mathematical function on the real-time embryo
approach. LabVIEW does contain extensive image processing tools as it is often used for
monitoring of process lines in industrial settings, so this proposal is likely buildable, but
beyond the scope of this project.
The obvious benefit of this system is the cost savings. Presently, vitrification requires
a large labour cost as an embryologist or highly skilled technician is needed for embryo
handling. This system eliminates the need for a high dexterity user and replaces it with
computer control. The infrastructure for the system is relatively inexpensive requiring
only signal generation equipment, an imaging system, and a computer for control. The
microfluidic chips are presently made in a cleanroom at significant cost. However, since
the devices only consist of a patterned chromium layer and two coatings on top, they
are easily manufactured at a large scale at a drastically lower cost. These chips would
be disposable in a clinical setting and switched between patients to ensure zero cross-
contamination.
Presently, the system reduces the skill required to complete the vitrification protocol,
yet the required time remains the same and cannot be reduced as it is dictated by
the embryo physiology leading to the protocol timings. However, the system can be
parallelized to reduce the time needed to freeze a large embryo sample. One easy way to
accomplish this would be to simply use multiple devices. The clinician could pipette a
single embryo onto each device, perform the protocol using a fully computer controlled
procedure, and then remove each embryo from the device. This would allow the device
design to remain simple, although if volume monitoring were needed, the imaging system
would be much larger to simultaneously monitor each device. Alternatively, the device
Chapter 3. On-Chip Embryo Vitrification 42Fig 1: Storage ring
embryo storage ring
Figure 3.12: A storage ring could be used to store multiple embryos each in their owndroplet. This would allow embryos to be loaded all at once, and then individually pro-cessed when needed.
design could be grown to handle multiple embryos at once. Embryos should still be
handled individually as with a human patient the embryo count can be very low (¡10),
so a storage system could be used to allow the user to load all the embryos at once,
then individually take each embryo from the storage system and through the vitrification
protocol.
A key addition necessary in scaling this technology up is an automated tool for trans-
ferring the embryo from the digital microfluidic device to the liquid nitrogen. This is
essential, as it would first eliminate variability and error from the user manually pipetting
and transferring the embryo off the device and plunging it into the liquid nitrogen. The
timing during the high concentration portion of the protocol is very critical so any added
time can have a significant effect. Also, in terms of cost reduction, the user presently
needs to load the embryos on the device and then return to transfer them into the liquid
nitrogen. If the user did not need to be present for this task the time savings would be
significantly larger. One method proposed would be to build a device that could be en-
tirely plunged into liquid nitrogen. This would completely eliminate the issue of embryo
Chapter 3. On-Chip Embryo Vitrification 43
extraction but posses new problems. First, by freezing a larger device the heat transfer
will be significantly slower in both freezing and thawing. This is known to decrease the
survivability of the embryos. Secondly, after the embryo has been frozen it needs to go
through a thawing protocol and be extractable. Reconnecting a thawing device is not
easily possible and if attempted would require a large amount of robotic tools. Overall,
freezing the device seems initially attractive, due to its simplicity in freezing, but makes
the task of thawing extremely difficult.
An alternate method would be to automate the entire embryo extraction procedure
and then robotically transfer the embryo into liquid nitrogen. The simplest method would
be to use a computer controlled robotic pipette and syringe pump to mimic what the
user is presently doing. However, this requires significant robotics and image processing
to achieve so ideally a solution easier to implement would be desired.
Another approach would be to add to the microfluidic system. Since droplets can be
moved inside and outside of the top ITO slide depending on the coverage of the ITO slide
on the exit electrode, the droplet could first be removed from the closed device. Secondly,
electrodes could be continued to be patterned on the open section of the device. Open
digital microfluidics is a commonly used approach although it suffers from evaporation
problems and the inability to split droplets. The embryo droplet could be moved on the
open device onto a detachable section of electrodes. This section of the device could
then be removed and plunged in liquid nitrogen. This is illustrated in Figure 3.13. This
method of transferring from a closed to open digital microfluidic platform has recently
been used to integrate a digital microfluidic chip with a mass spectrometer [57].
This proposed mechanism raises significant feasibility questions yet if the microfluidics
are possible, would produce a simple device that fully automates the process. Addition-
ally, in the design of the detachable device carrier, a flexible substrates can also be used
to guide droplet movement and open new design possibilities [58].
A silicon oil bath was used on the device to aid in droplet motion by lowering the
Chapter 3. On-Chip Embryo Vitrification 44
Fig 2: Open DMF
(a) (b) (c)
(d) (e)
top ITO slide
detachable section
Figure 3.13: Schematic showing droplet transfer to removable freezing device. (a) Embryoinitially inside device, (b) droplet moved outside of closed structure, (c) droplet movedusing open digital microfluidic electrodes, (d) droplet transported onto freezing device,and (e) freezing device removed and directly plunged into liquid nitrogen.
Chapter 3. On-Chip Embryo Vitrification 45
Fig 3: Oil floating
(a) (b)
droplet in air droplet in oil
Figure 3.14: (a) Droplet is in contact with substrate in air and (b) floating over a smalloil layer when in an oil bath.
effective friction of the system [59]. For basic liquids, like deionized water, this is not
normally needed but can still aid in high speed applications. For vitrification mediums,
which were of higher viscosity, the oil bath was found to be necessary for strong, reliable
motion. However, this also has a second set of benefits particularly useful for critical
biological applications. First, evaporation is eliminated from the device. This is critical
as if the concentration of vitrification mediums is altered the entire process becomes
unusable. Typically when the procedure is performed manually a mineral oil bath is
used to cover each well of liquid. Secondly, electrode contamination is reduced. Without
oil small amounts of contaminates have been found to remain on electrodes after use
making multiple uses of a single device not advisable [60]. With an oil bath, the droplet
is actually floating in the device with a small oil layer above and below the device,
eliminating residue [55].
Since an electric field is present in the device to drive the droplets, this also raises
the concern that this may pose negative effects on the health of the embryo. However, it
has been shown that for small electrodes and low driving frequencies that no significant
damage is done to mammalian cell lines [61]. Additionally, architectures requiring much
lower voltages are have been developed which may result in even lower cellular stress due
to electric effects [62].
Chapter 3. On-Chip Embryo Vitrification 46
3.10 Conclusion
This chapter described a digital microfluidic device for automated embryo processing for
vitrification applications. The results demonstrated that a high embryo development rate
can be achieved using the automated approach. Advantages of this approach, compared
to manual operation and channel-based microfluidic vitrification, include automated op-
eration, cryoprotectant concentration gradient generation, and feasibility of loading and
retrieval of embryos. The device permits one to readily modify/test vitrification proto-
cols with significant reduction in labor costs. Further development can possibly facilitate
new vitrification protocol development and clinical IVF practice.
Chapter 4
Partially Filled Electrodes
4.1 Introduction
Digital microfluidic devices, as a platform technology, enable programmed manipulation
of small droplets on arrays of microelectrodes [63, 64] for performing tasks such as PCR,
cell culture, and immunoassays [27–29]. In digital microfluidic devices, conductive or
polar droplets are moved under the effect of electrodynamic forces. These forces are
generated by the electric field induced from the energized electrode beneath the droplet
[65].
To increase capabilities of digital microfluidic devices, researchers have begun inte-
grating additional elements within electrodes such as impedance spectroscopy [66], elec-
trophoresis electrodes for particle separation [67], absorbance detection windows [53],
heaters [68], field effect transistor-based biosensors [69] and cell culture patches [27].
As these elements are added, electrode area for droplet actuation is reduced, and force
generation becomes weaker. Electrodes cannot simply be scaled larger to compensate
for this lower actuation force since this would also increase the droplet volume thereby
compromising miniaturizing advantages and increasing device foot print. Therefore, un-
derstanding the effect of reducing electrode area on generated forces and droplet speed
47
Chapter 4. Partially Filled Electrodes 48
(a) (b)
(c)
partially filled electrode
filled/opaque chromium region
removed/transparent region
Figure 4.1: (a) Schematic of partially filled electrodes providing space for additional on-chip tools or as a window for imaging. (b) Example designs of partially filled electrodeconfigurations considered. Electrodes are 1mm x 1mm (c) Image of droplet on series ofpartially filled electrodes.
is necessary.
Vitrification in previous chapters was done under reflective imaging, while transmis-
sion microscopy is the most often used imaging platform. Therefore, this chapter presents
a detailed study evaluating different partially filled electrode designs and suggesting de-
signs that combine a high actuation force with a large reduction in electrode area that
permits integration of large elements within the electrode. As a sample application, a
non-ITO, partially filled Cr electrode design that permits the imaging of droplet contents
using standard transmission microscopy is presented.
4.2 Modelling and Simulation
To quantify the effect of removing electrode area, actuation forces on different designs of
partially filled electrodes were simulated using finite element analysis (COMSOL Multi-
physics). Relevant dimensions used in simulation are summarized in Table 4.1. Droplet
was modelled as a non-deformable lossy dielectric and hence, the conservation of charge
Chapter 4. Partially Filled Electrodes 49
Table 4.1: Parameters used in numerical simulations.
summarized in Table I. Droplet was modelled as a non-deformable lossy dielectric and hence, the conservation ofcharge [Eq. (1)] and Laplace [Eq. (2)] applies12
rðrrVÞ ¼ 0; (1)
rðerVÞ ¼ 0; (2)
where e ¼ ere0 (er is the relative permittivity of the medium,and e0 is the permittivity of free space), r is the electricalconductivity of the droplet, and V is the electric potential.Actuation forces are then calculated by integrating theMaxwell stress tensor over the surface of the droplet, assum-ing negligible magnetic fields,13 according to
F ¼ð
ST $ nds; (3)
Tij ¼ eðEiEj % 0:5 dijE2Þ; (4)
where Tij is the Maxwell stress tensor. Actuation forces werecalculated over a series of droplet locations along its actua-tion path. Since the integration of the Maxwell stress tensoris dependent on mesh geometry and density, a uniform sweptmesh is used to allow for a fine mesh on the droplet leadingand trailing faces without considerably increasing the totalnumber of elements, Fig. 2(a). Swept layers are distributedin the vertical direction following a geometric sequence sothat layers are denser near the underlying dielectric layer toavoid a large jump in element size at the droplet and dielec-tric interface.
Fig. 2(b) shows force curves as a function of dropletposition. This approach enables quantifying the quality ofany arbitrary electrode design. Forces on the leading surfaceof the droplet are dominant except at the end of dropletmotion when the trailing surface produces a reverse force onthe droplet, which leads to droplet motion stopping on thecenter of the electrode. Position is defined as the distancefrom the left edge of the electrode to the leading edge of thedroplet.
The goal of simulations was to address two questionsin the design of partially filled electrodes: how theremoved electrode area affects actuation forces andwhere electrode removal is most critical to force genera-tion. To answer the first question, a horizontally stripedelectrode design with variable strip width was simulated,and the relationship between electrodynamic force andelectrode fill percentage was determined [Fig. 2(c)].Force was found to have a linear relationship with fillpercentage, which is expected as the majority of force isgenerated at the advancing three phase contact line of thedroplet. In these simulations, the length of the contact
FIG. 2. Simulation results. (a) Droplethalf space mesh and swept uniformmesh on droplet surfaces. (b) Actuationforce on droplet for conventional elec-trode on leading and trailing dropletsurfaces. Reverse actuation force isgenerated on trailing droplet surface asbackward interface begins to moveonto electrode. (c) Induced forcesincrease linearly with electrode fill ra-tio, which was changed by varying thewidth of the horizontal bars in the elec-trode. (d) Force is independent of verti-cal location of removed area fromelectrode. The leading edge of thedroplet is fixed at the midpoint of theelectrode for panels (c) and (d).
TABLE I. Parameters used in numerical simulation.
Parameter Value
Electrode dimensions 1& 1 mm2
Dielectric thickness 10 lm
Droplet radius 740 lm
Droplet contact radius 600 lm
Droplet height 140 lm
Droplet conductivity (r) 5.5& 10%6 S/m
Relative permittivity of droplet 80
Relative permittivity of air 1
Relative permittivity of dielectric 2.5
Actuation voltage 100 V
024103-2 Pyne et al. Appl. Phys. Lett. 103, 024103 (2013)
This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:142.150.190.39 On: Wed, 16 Oct 2013 02:31:57
(Eq. 4.1) and Laplace (Eq. 4.2) applies [70].
5(σ5 V ) = 0 (4.1)
5(ε5 V ) = 0 (4.2)
where ε = εrε0 (εr is the relative permittivity of the medium, and ε0 is the permittivity of
free space), σ is the electrical conductivity of the droplet, and V is the electric potential.
Actuation forces are then calculated by integrating the Maxwell stress tensor over the
surface of the droplet, assuming negligible magnetic fields [71], according to
F =
∫S
T · nds (4.3)
Tij = ε(EiEj − 0.5δijE2) (4.4)
where Tij is the Maxwell stress tensor. Actuation forces were calculated over a series of
droplet locations along its actuation path. Since the integration of the Maxwell stress
tensor is dependent on mesh geometry and density, a uniform, swept mesh is used to allow
for a fine mesh on the droplet leading and trailing faces without considerably increasing
the total number of elements, Fig. 4.2(a). Swept layers are distributed in the vertical
Chapter 4. Partially Filled Electrodes 50
direction following a geometric sequence so that layers are denser near the underlying
dielectric layer to avoid a large jump in element size at the droplet and dielectric interface.
Fig. 4.2(b) shows force curves as a function of droplet position. This approach enables
quantifying the quality of any arbitrary electrode design. Forces on the leading surface
of the droplet are dominant except at the end of droplet motion when the trailing surface
produces a reverse force on the droplet, which leads to droplet motion stopping on the
center of the electrode. Position is defined as the distance from the left edge of the
electrode to the leading edge of the droplet.
The goal of simulations was to address two questions in the design of partially filled
electrodes: how the removed electrode area affects actuation forces; and where electrode
removal is most critical to force generation. To answer the first question, a horizontally
striped electrode design with variable strip width was simulated, and the relationship
between electrodynamic force and electrode fill percentage was determined (Fig. 4.2(c)).
Force was found to have a linear relationship with fill percentage, which is expected as
the majority of force is generated at the advancing three phase contact line of the droplet.
In these simulations the length of the contact line in contact with the electrode increases
approximately linearly as the fill percentage is increased (i.e. as the horizontal strips are
widened), which leads to the linear increase in force.
To answer the second question of assessing the location effect of electrode area removal
on the induced force, an electrode with a single horizontal window was studied. The
distance between the window and the centerline of the electrode was then varied to test
the position dependence in the perpendicular axis (Fig. 4.2(d)). Vertical position of
the removed portion of the electrode was found to have no effect on force generation.
Since most of the actuation force is generated on the contact line, this result can be
explained by considering the amount of force lost in the horizontal window. As the
window moves farther from the centerline and the contact line curves, a larger length of
it is placed inside the window. However, the generated force per length is normal to the
Chapter 4. Partially Filled Electrodes 51
(b)
Forc
e (µ
N)
Fill Ratio (%) 30 40 50 60 70 80 90 1
1.5
2.0
2.5
3.0
3.5
Varied bars’ heights
Forc
e (µ
N)
Horizontal Bar Gap Position (µm) 50 100 150 200 250 300 350 400
2.5
2.7
2.9
3.1
3.3
3.5
3.7
3.9 4.0
Gap Position
(c) (d)
(a)
Forc
e (µ
N)
Position (µm)
Total Leading Surface Trailing Surface
0 200 400 600 800 1000 1200 -1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
4
Figure 4.2: Simulation results. (a) Droplet half space mesh and swept uniform mesh ondroplet surfaces. (b) Actuation force on droplet for conventional electrode on leadingand trailing droplet surfaces. Reverse actuation force is generated on trailing dropletsurface as backward interface begins to move onto electrode. (c) Induced forces increaselinearly with electrode fill ratio, which was changed by varying the width of the horizontalbars in the electrode. (d) Force is independent of vertical location of removed area fromelectrode. The leading edge of the droplet is fixed at the midpoint of the electrode forpanels (c) and (d).
Chapter 4. Partially Filled Electrodes 52
droplet surface; therefore, the component of the actuation force parallel to the centerline is
proportional to the vertical projection of the contact length. Since this projection remains
constant as the window moves away from the centerline, there is no force dependence
on the position of the removed portion. These results lead to the conclusion that force
generation is dominated by total electrode area and at any particular position, by the
vertical projection of the three phase contact line in contact with the filled area of the
electrode.
In addition to its effect on the generated force on the droplet, filled areas of the
electrode should be distributed in a manner that does not compromise the initial pulling
force at the beginning of the motion where generated forces are at its lowest. This is
critical to produce a large enough force to induce motion and overcome line pinning.
Large initial pulling forces can be achieved by making sure that the entrance area of the
electrode is always completely filled as a crescent that matches the droplets leading edge,
as shown in Fig. 4.3. Removal of a portion of the electrode, in the form of horizontal
stripes or any other form, can then follow to allow for the integration of various elements
into the electrode. This shape guarantees that maximum force is applied to the advancing
contact line at the crucial stage of initiating droplet motion. This filled crescent is
mirrored at the other side of the electrode to provide strong droplet motion in both
directions, but it can be removed if droplet motion is desired only in one direction allowing
for an even larger area for the integration of other device elements.
4.3 Experimental Evaluation
To experimentally evaluate the effect of reducing the energized electrode area, the max-
imum droplet actuation frequency (i.e., how many electrodes can the droplet cross per
second) on partially filled and solid electrodes is compared. Droplet motion was con-
trolled using a custom LabVIEW program controlling an array of 24 relays connected to
Chapter 4. Partially Filled Electrodes 53
0
10
20
30
40
50
60
Max
imum
Act
uatio
n Fr
eque
ncy
(Hz)
Crescent design
Horizontal bars
Entrance/exit fill increase initial force
Forc
e (µ
N)
Position (µm)
Improved Entrance/exit Design Horizontal Bars
0 200 400 600 800 1000 1200 0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8 (a) (b)
100 µm 75 µm 40 µm 100 % Filled
Horizontal Bar Width (d)
unfilled
d
Figure 4.3: Simulation results: electrode design with crescent-like filled areas at theentrance and exit of the electrode produces increased force at beginning and end ofdroplet motion to create initial force to ensure droplet motion is generated.
a high voltage amplifier (Trek PZD350A, Medina NY). Motion was then recorded using
a CCD camera (Basler acA1300-30gc, Exton PA) connected to a motorized zoom lens
(Navitar 12X Body Tubes, Rochester NY). As expected, partially filled electrodes showed
a decrease in maximum actuation frequency and droplet motion speed, since only seg-
ments of the contact line are exposed to high electric fields [65]. Nevertheless, actuation
frequencies of over 10 electrodes per second were achieved on devices with partially filled
electrodes at a low electrode fill area of 40%, Fig. 4.4. This speed is sufficient for many
biology and clinical applications of digital microfluidic devices [72].
Removed areas on partially filled electrodes can be used for integrating other elements
(e.g. heaters, detectors and hydrophilic patches) into the electrodes; furthermore, they
can be useful for on-chip imaging of droplet contents using transmission microscopy.
Transmission microscopy imaging, particularly on inverted microscopes is a standard
platform used in biology labs and clinics. Since completely filled, non-transparent elec-
trodes made of metallic materials (e.g., Cr or gold) are not compatible with transmission
microscopy imaging, ITO is typically used to construct electrodes [24, 28, 68]. However,
Chapter 4. Partially Filled Electrodes 54
0
10
20
30
40
50
60
Max
imum
Act
uatio
n Fr
eque
ncy
(Hz)
Crescent design
Horizontal bars
Entrance/exit fill increase initial force
Forc
e (µ
N)
Position (µm)
Improved Entrance/exit Design Horizontal Bars
0 200 400 600 800 1000 1200 0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8 (a) (b)
100 µm 75 µm 40 µm 100 % Filled
Horizontal Bar Width (d)
unfilled
d
Figure 4.4: Experimental comparison of electrode designs. In experiments, maximumdroplet actuation frequency was measured at different fill percentages for normal andimproved designs. Number of horizontal bars in the electrode was kept constant, thusreducing the bar width, reduces electrode fill percentage. Please note that reduction inmaximum actuation frequency is almost proportional to the reduction in the bar widthsimilar to the force reduction simulations. Experiments were conducted with deionizedwater at 75 V rms and 15 kHz by actuating droplet back and forth across a series of 5electrodes at increasing speed until droplet motion could not keep up with actuation.
Chapter 4. Partially Filled Electrodes 55
ITO is more expensive in materials and fabrication than metallic electrodes, and the
invisible ITO electrodes can pose challenges in device debugging. In comparison, par-
tially filled metallic electrodes can be easier to construct while being compatible with
transmission microscopy imaging.
Therefore, intentionally constructed electrodes using metallic materials were used in
this study to demonstrate their compatibility with standard transmission microscopy
imaging. Partially filled Cr electrodes were used to actuate single mouse embryos. Mor-
phology of the mouse embryos on partially filled electrodes was imaged under transmis-
sion differential interference contrast (DIC) microscopy, which revealed detailed embryo
morphology and cell structures (Fig. 4.5(a)). In applications, if the droplet medium is
relatively viscous, the cells position would remain relatively constant within the droplet.
In this case, a set of partially filled electrodes negative of each other can be used to ensure
that the cell can be imaged across these electrodes. If the cells position does not remain
constant with a less viscous droplet, the droplet can be easily actuated back and forth
on partially filled electrodes until the cell rests on a non-filled region. Since it is viable to
remove half of the electrode area without affecting generated forces significantly, a few
actuations is typically sufficient to achieve cell imaging with this technique.
Also droplets containing human red blood cells (RBCs) were actuated using partially
filled Cr electrodes. PBS droplets containing RBCs were mixed with differing amounts of
DI water (i.e. varied osmolarity). RBC viability as a function of medium osmolarity was
measured on chip. Lysed RBCs were easily identifiable on our partially-filled electrodes
under transmission microscopy, Fig. 4.5(b). The non-filled section on a single electrode
was large enough to sample a significant portion of the population to gather meaningful
statistics.
Chapter 4. Partially Filled Electrodes 56
0 10 20 30 40 50 60 70 80 90
100
75 125 175 225 275
% R
BC
s ly
sed
Osmolarity (mOsm/L)
(b)
intact cell
lysed cell 100µm
!!
!!
100 µm
(a)
Figure 4.5: (a) Single mouse embryo morphology on partially filled Cr electrodes. Upperimage uses bright field transmission DIC imaging showing superior detail compared toreflection microscopy imaging used in lower image. (b) RBC viability measured withdifferent osmolarity by counting cells through unfilled regions on chip. N = 150 600 foreach osmolarity point.
4.4 Conclusion
In summary, this chapter investigated, numerically with preliminary experimental verifi-
cation, the effect of removing sections of electrodes in digital microfluidic devices on the
generated electrodynamic forces and the maximum achievable droplet speed. Generated
electrodynamic forces were found to be linearly dependent on electrode fill percentage
and independent of position of the horizontal non-filled areas. To maintain high initial
pull for the droplet, entrance and exit areas of the electrodes were left completely filled.
As application examples, partially filled metallic electrodes were constructed and their
compatibility with standard transmission microscopy imaging demonstrated. These re-
sults are meaningful for guiding the design of digital microfluidic devices that require
the integration of other elements, such as detectors, heaters, and cell culture patches, on
electrodes.
Chapter 5
Summary
5.1 Conclusions
In this work, a digital microfluidic platform was first built, including device fabrica-
tion, an imaging system, a high voltage control system, and a LabVIEW interface. The
main purpose of this system was to automate the cryopreservation process of vitrifi-
cation, which can achieve a higher survival rate then conventional methods yet has a
high labour cost requirement. This system automated the liquid handling and timing
tasks in embryo vitrification. Technical advantages of this approach, compared to man-
ual operation and channel-based microfluidic vitrification, include automated operation,
cryoprotectant concentration gradient generation, and feasibility of loading and retrieval
of embryos. The device permits researchers to readily change/test protocols. Significant
labor costs were also reduced by eliminating the need for highly skilled operators. The
device showed cell survival and development rates of 77% and 90%, respectively, which
are comparable to the control groups that were manually processed.
As a secondary task, the effect of removing sections of electrodes in digital microflu-
idic devices on the generated electrodynamic forces and the maximum achievable droplet
speed was studied through simulation and experimentation. Generated electrodynamic
57
Chapter 5. Summary 58
forces were found to be linearly dependent on electrode fill percentage and independent
of position of the horizontal non-filled areas. To maintain high initial pull for the droplet,
entrance and exit areas of the electrodes were left completely filled. As application ex-
amples, partially filled metallic electrodes were constructed and their compatibility with
standard transmission microscopy imaging demonstrated. These results are meaningful
for guiding the design of digital microfluidic devices that require the integration of other
elements, such as detectors, heaters, and cell culture patches, on electrodes.
5.2 Future Directions
The following are examples of future works that can be undertaken:
1. Testing the physiological effects of Pluronic additives on mammalian embryos and
whether it increases droplet movement enough to allow the use of conventional
embryo culture mediums which contain serum.
2. Development of a system to transfer an embryo from the digital microfluidic device
into liquid nitrogen by either using a detachable open digital microfluidic segment
or by robotically pipetting the embryo containing droplet off of the device.
3. Development of an image processing algorithm to monitor embryo volume in real-
time to provide physiological feedback and allow the development of closed loop
vitrification protocols.
4. Establishment of an on-chip storage system to decrease required technician work-
load in vitrifying a large sample of embryos. Integration of a warming platform
may also be required to provide the culture conditions to allow embryos to safely
remain on the device for longer periods of time.
5. Integration of capacitance measurements to detect droplet locations and enable
closed loop droplet motion protocols.
Chapter 5. Summary 59
6. Design of a system to automate the thawing portion of the cryopreservation proce-
dure.
7. Testing of the current system’s ability to vitrify other biological samples of interest
such as cells, tissues and sperms.
Appendix A
Detailed Fabrication Recipe
Fabrication protocol for digital microfluidics chips:
1. Process is started with precoated 100 nm chromium glass slides (50 x 75 mm). If
these are not available chromium can be deposited by metal evaporation. However,
using precoated slides was found to save significant amounts of time and was even
cost effective, as metal evaporation requires 3-4 hours.
2. Slides are cleaned by rinsing in acetone, isopropanol, and de-ionized water (DI
water). Wafer handling tweezers are used to hold slides and rinsing is done into
a beaker so that solvents can be properly disposed of after cleaning. Acetone and
isopropanol are kept in squeeze bottles while DI water is available through the wet
bench taps by a DI water generation system in the cleanroom. The isopropanol
is used to remove residue left by the acetone, and the DI water provides the final
clean. Ensure that while rinsing the slide is repositioned in each step to reach the
area of the substrate underneath the tweezers.
3. Substrates are blown dry with a nitrogen gun. To ensure complete dehydration,
slides are placed on a hotplate at 115◦C for 5 minutes. Slides are then removed
and given another 5 minutes to cool to room temperature.
60
Appendix A. Detailed Fabrication Recipe 61
4. Slides were then liquid primed with P-20 (20% HMDS) primer. Slides are placed
on a chuck in the spin coater one at a time and the vacuum is engaged, which
holds the slide to the chuck. It is a good idea to first test the vacuum strength
by performing a dry run with a disposable glass slide. This avoids the possibility
of damaging one of the more valuable cleaned chromium slides. Primer is applied
to the entire wafer with a soft pipette. A reminder that a new pipette should be
used for each medium to avoid cross contamination. Allow primer to remain for
10 seconds, and then spin at 3000 rpm for 30s with an acceleration setting of 8.
When opening the spin coater lid after coating, it is a good idea to hold a clean
room tissue over the sample to protect it from any drips off of the lid.
5. Keeping the slide in the spin coater, photoresist S1811 is dispensed over the slide
using a new pipette. Photoresist should be poured into a small beaker and dispensed
from this beaker to avoid contaminating the supply bottle. Ensure that ample
photoresist is used to coat the entire area of the slide. Since the photoresist has
a high viscosity, uncoated areas may leave streaks during spin coating. Spin using
the same settings of 3000 rpm for 30s with an acceleration setting of 8. Be extra
careful when lifting the lid after this step as due to the large amount of photoresist
used there is a higher chance of drips.
6. Samples are then soft baked at 115◦C for 2 minutes on a hot plate to remove some of
the solvents inside the photoresist. This partially hardens the photoresist coating.
7. Samples are now ready for UV exposure using the mask aligner. Slides should be
brought to the mask aligner station one at a time in case there is any stray UV
exposure while the aligner is running. Firstly, ensure that the nitrogen, vacuum,
and compressed air lines to the aligner are open. This is critical as they provide
cooling to the UV bulb, which can overheat and crack without it. The mask aligner
and UV bulb can then be turned on. Wait for approximately 10 minutes until the
Appendix A. Detailed Fabrication Recipe 62
UV bulb has warmed up, it will display its power usage once it is warm. The flood
exposure setting is used which simply exposes the entire sample without vacuuming
the sample against a glass plate or glass photomask. This is the quickest setting to
use, compared to hard or soft contact, yet produces the lowest resolution. However,
it is sufficient for this application as the smallest feature size is 20 µm. Once the
flood exposure setting is engaged, the exposure time is set to 10 s. This slightly
over exposes the sample to ensure that no short circuits are present as the failure
case for over exposing is only slightly slower motion, which is much better then
the failure case for under exposure, which is short circuits producing an unusable
device. Place a single slide into the mask aligner and lay the photomask directly on
top of it. Align the photomask with the edges of the glass slide. Alignment is not
very critical as the only concern is that the entire design is on the slide and that
it is roughly parallel with the slide edge. Exposure can now be performed. While
exposing, make sure to not look directly at the machine, as the UV light can be
harmful. This process is then repeated for each slide, after which they should be
returned to the wet bench.
8. To prepare for development and etching, it is best to prepare the mediums in a line
across the wet bench to produce an organized process flow. This is more important
as multiple slides can be processed in parallel and establishing a process line can
help reduce the chances of error. The exposed samples are grouped on the far
right side of the wet bench. The DI water tap is on the left side of the referenced
wet bench so the process works from right to left in this description. Moving left,
MF-321 developer is poured into a wide, low beaker. Enough liquid should be
dispensed to cover the slides by approximately 1 cm of liquid. Next a hot plate is
placed at 115◦C. A cleanroom tissue is placed to the left of the hot plate to act
as a cooling station. Chromium etchant, CR-4 is dispensed in a similar wide, low
beaker, also with enough liquid to cover slides by approximately 1 cm of liquid. A
Appendix A. Detailed Fabrication Recipe 63
clean room tissue is then the last item and acts as a finishing area for the etched
devices. One sample should only be placed in each bath at a time to prevent slides
from scratching each other.
9. Slide is next placed in developing solution for 2-4 minutes. Some small agitation
can be done to speed up the process by either swirling the solution with the beaker,
or lifting one end of the slide up and down with a pair of tweezers. Development
is complete when photoresist is no longer clinging, or dissolving off of the surface
of the slide. The risk of over development is much smaller then the risk of under
development for these large feature sizes. This makes it a good idea to keep the
slide in the development solution for an additional 30 seconds after development is
judged to be complete.
10. Development solution is rinsed off with DI water. Again, make sure to adjust the
tweezers position to rinse the area under the initial tweezers placement. Dry the
sample with the nitrogen gun. Hard bake the slide at 115◦C for 1 minute to remove
the last remaining solvents in the photoresist. Place the sample on the cooling
station to the left of the hotplate and allow cooling to room temperature.
11. Etching is performed by placing the slide in the chromium etchant solution. Slight
agitation can be performed in the same manner as during development. Etching
requires approximately 2 minutes and it is also suggested to keep the sample in
the solution for an additional 30 seconds after etching is deemed to be complete.
During etching the chromium not protected by photoresist will dissolve so that
the final device will be visible. Rinse in DI water and dry with the nitrogen gun.
Samples are then placed in the finishing area.
12. Devices should now be inspected under microscope to ensure that lithography pat-
terns look good and no short circuits are present. If the devices look accurate, the
remaining devices can be etched without inspection.
Appendix A. Detailed Fabrication Recipe 64
13. Parylene C coating is performed using a dedicated chemical vapor deposition ma-
chine. Samples are placed on a rotating rack inside the vacuum chamber. Ap-
proximately 10 samples can fit in the machine during a run. To prepare samples,
fold a piece of tape over the long edge of each slide to cover the edge electrodes.
This prevents parylene from depositing on these areas, which makes it quicker to
assemble devices, as parylene needs to be removed in these areas to ensure a strong
connection to the edge board connector. Make sure that the tape is pressed well
on the glass slide as parylene is a conformal coating and will coat any exposed
surfaces. Parylene C is initially in solid pellet form and the deposition thickness
determined by the weight of parylene loaded into the machine. Parylene is loading
on an aluminum foil boat that is placed in an initial heating chamber at the bottom
of the machine. 3 grams of parylene was found to produce an appropriate coating
of approximately 1.5 µm. The chamber can now be closed and deposition initiated.
If the vacuum seal is strong enough, the machine will automatically heat and de-
posit the parylene inside the vacuum chamber. If proper vacuum is not reached,
the chamber needs to be vented and seals checked for debris. Deposition occurs at
room temperature so there is no risk of melting any materials on the sample. After
deposition is complete, solvents should not be used on the devices.
14. Teflon solution is prepared by placing 0.25 grams of Teflon AF 1600 into a glass
vial with 25 grams of FC-40. This produces a 1% Teflon solution. The vial is then
placed in an oven or hotplate at 80◦C for several hours until all Teflon has dissolved
in the solution. This solution can be prepared beforehand.
15. The final Teflon coating is then applied by spin coating. Samples are placed back
in the spin coater and the Teflon solution dispensed with a soft pipette. Samples
are spun at 2000 rpm for 1 minute. Lastly, solvents are removed by hard baking
devices at 160◦C for 10 minutes. Slides can then be allowed to cool and are now
Appendix A. Detailed Fabrication Recipe 65
complete.
16. Precoated indium tin oxide (ITO) slides are used for the top ground plate. These
slides are also coated in Teflon using the same spin coating technique.
Appendix B
Selected LabVIEW Code
66
Appendix B. Selected LabVIEW Code 67
Fig
ure
B.1
:C
amer
aco
mm
unic
atio
nis
open
edusi
ng
IMA
Qco
mm
ands.
Som
ein
itia
lop
tion
sar
ew
ritt
ensu
chas
pac
ket
size
,ga
mm
aco
ntr
ol,
and
exp
osure
tim
e.T
his
cam
era
feed
isfe
din
toth
eim
age
captu
ring
loop
.
Appendix B. Selected LabVIEW Code 68
Fig
ure
B.2
:T
osi
mplify
cam
era
adju
stm
ents
,co
lor
bal
anci
ng
isdon
eau
tom
atia
lly
usi
ng
asi
ngl
ebutt
onon
the
fron
tpan
el.
When
the
’Auto
Bal
ance
’is
pre
ssed
the
gain
and
whit
ebal
ance
are
auto
mat
ical
lyad
just
edusi
ng
funct
ions
onth
eca
mer
a.T
hes
ese
ttin
gsca
nb
ere
bal
ance
dduri
ng
exp
erim
enta
tion
ifligh
ting
orsa
mple
condit
ions
chan
ge.
Appendix B. Selected LabVIEW Code 69
Fig
ure
B.3
:In
side
the
mai
nim
agin
glo
op,
fram
esar
eta
ken
from
the
cam
era
and
dis
pla
yed
onan
imag
ing
win
dow
inth
efr
ont
pan
el.
When
the
’Rec
ord’
butt
onis
init
ially
pre
ssed
,a
AV
Ico
nta
iner
file
usi
ng
anM
PE
Gco
mpre
ssor
iscr
eate
dusi
ng
the
curr
ent
tim
esta
mp
asth
efile
nam
eto
ensu
reth
atth
ere
are
no
file
nam
eco
nflic
ts.
Aft
erth
isin
itia
lly
loop
,if
the
’Rec
ord’
butt
onis
still
pre
ssed
,in
div
idual
lyfr
ames
are
added
toth
eA
VI
conta
iner
file
.A
tim
eris
also
use
dto
dis
pla
yth
efr
ame
rate
onth
efr
ont
pan
el.
Appendix B. Selected LabVIEW Code 70
Figure B.4: Communication with the signal generator is achieved over USB communi-cation. Standard VISA commands are used to set the function, voltage, and frequencyvalues. Communication is sent to the signal generator whenever either the voltage, orfrequency control values on the front panel are changed.
Appendix B. Selected LabVIEW Code 71
Figure B.5: Macro control is initiated by reading in excel files containing the list ofdroplet actuations for each command. These files are named to match their function andeach button on the front panel corresponds to an individual excel file. Once the file isopened it is fed into the main droplet control loop.
Appendix B. Selected LabVIEW Code 72
Fig
ure
B.6
:In
the
mai
ndro
ple
tco
ntr
ollo
op,
refe
rence
sto
each
elec
trode
butt
onar
efirs
tbuild
into
anar
ray
soth
atth
eyar
ein
dex
ed.
Eac
hline
ofex
cel
file
isth
enre
adan
dth
eco
rres
pon
din
gel
ectr
odes
turn
edon
.O
pti
ons
are
avai
lable
tole
ave
the
rese
rvoi
rel
ectr
odes
onat
all
tim
esto
ensu
reth
atth
ela
rge
rese
rvoi
rdro
ple
tis
not
mov
ed.
The
tim
eb
etw
een
actu
atio
ns
can
also
be
set
usi
ng
afr
ont
pan
elco
ntr
ol.
Com
munic
atio
nw
ith
anL
CR
met
erw
asal
soin
itia
ted
tose
nse
dro
ple
tlo
cati
ons
afte
rea
chst
ep.
How
ever
,th
edro
ple
tm
otio
nw
asfo
und
tob
eve
ryro
bust
,m
akin
gth
eex
tra
com
ple
xit
yof
dro
ple
tse
nsi
ng
not
nec
essa
ry.
Bibliography
[1] Alan Trounson and Linda Mohr. “Human pregnancy following cryopreservation,
thawing and transfer of an eight-cell embryo”. In: Nature 305.5936 (Oct. 1983),
pp. 707–709. issn: 0028-0836. doi: 10.1038/305707a0. url: http://www.natur
e.com/nature/journal/v305/n5936/abs/305707a0.htmlhttp://www.nature.c
om/doifinder/10.1038/305707a0.
[2] B E Reubinoff et al. “Effective cryopreservation of human embryonic stem cells
by the open pulled straw vitrification method.” In: Human reproduction (Oxford,
England) 16.10 (Oct. 2001), pp. 2187–94. issn: 0268-1161. url: http://www.ncb
i.nlm.nih.gov/pubmed/11574514.
[3] CMO Medeiros et al. “Current status of sperm cryopreservation: why isn’t it bet-
ter?” In: Theriogenology 01 (2002), pp. 327–344. url: http://medcontent.metap
ress.com/index/A65RM03P4874243N.pdfhttp://www.sciencedirect.com/scie
nce/article/pii/S0093691X01006744.
[4] Amir Arav et al. “New trends in gamete’s cryopreservation”. In: Molecular and
Cellular Endocrinology 187.1-2 (Feb. 2002), pp. 77–81. issn: 03037207. doi: 10.1
016/S0303-7207(01)00700-6. url: http://linkinghub.elsevier.com/retrie
ve/pii/S0303720701007006.
[5] Wai Hung Tsang and King L Chow. “Cryopreservation of mammalian embryos:
Advancement of putting life on hold.” In: Birth defects research. Part C, Embryo
73
BIBLIOGRAPHY 74
today : reviews 90.3 (Sept. 2010), pp. 163–75. issn: 1542-9768. doi: 10.1002/bdrc.
20186. url: http://www.ncbi.nlm.nih.gov/pubmed/20860056.
[6] K Oktay et al. “Cryopreservation of immature human oocytes and ovarian tissues:
an emerging technology?” In: Fertility and sterility 69.1 (1998), pp. 1–7. url: http:
//www.sciencedirect.com/science/article/pii/S0015028297002070.
[7] Murat Sonmezer and Kutluk Oktay. “Fertility preservation in female patients.” In:
Human reproduction update 10.3 (2004), pp. 251–66. issn: 1355-4786. doi: 10.10
93/humupd/dmh021. url: http://www.ncbi.nlm.nih.gov/pubmed/15140872.
[8] Stephanie J Lee et al. “American Society of Clinical Oncology recommendations on
fertility preservation in cancer patients.” In: Journal of clinical oncology : official
journal of the American Society of Clinical Oncology 24.18 (June 2006), pp. 2917–
31. issn: 1527-7755. doi: 10.1200/JCO.2006.06.5888. url: http://www.ncbi.n
lm.nih.gov/pubmed/16651642.
[9] W Hamish B Wallace, Richard a Anderson, and D Stewart Irvine. “Fertility preser-
vation for young patients with cancer: who is at risk and what can be offered?” In:
The lancet oncology 6.4 (Apr. 2005), pp. 209–18. issn: 1470-2045. doi: 10.1016/S1
470-2045(05)70092-9. url: http://www.ncbi.nlm.nih.gov/pubmed/15811616.
[10] Ryan Riggs et al. “Does storage time influence postthaw survival and pregnancy
outcome? An analysis of 11,768 cryopreserved human embryos.” In: Fertility and
sterility 93.1 (Jan. 2010), pp. 109–15. issn: 1556-5653. doi: 10.1016/j.fertnste
rt.2008.09.084. url: http://www.ncbi.nlm.nih.gov/pubmed/19027110.
[11] Donna Dowling-Lacey et al. “Live birth from a frozen-thawed pronuclear stage
embryo almost 20 years after its cryopreservation.” In: Fertility and sterility 95.3
(Mar. 2011), 1120.e1–3. issn: 1556-5653. doi: 10.1016/j.fertnstert.2010.08.
056. url: http://www.ncbi.nlm.nih.gov/pubmed/20887984.
BIBLIOGRAPHY 75
[12] David E Pegg. “The relevance of ice crystal formation for the cryopreservation of
tissues and organs.” In: Cryobiology 60.3 Suppl (July 2010), S36–44. issn: 1090-
2392. doi: 10.1016/j.cryobiol.2010.02.003. url: http://www.ncbi.nlm.ni
h.gov/pubmed/20159009.
[13] DG Whittingham. “Survival of mouse embryos after freezing and thawing”. In:
Nature 233.5315 (1971), pp. 125–6. url: http://www.nature.com/nature/journ
al/v233/n5315/abs/233125a0.html.
[14] Joseph Saragusty and Amir Arav. “Current progress in oocyte and embryo cryop-
reservation by slow freezing and vitrification.” In: Reproduction (Cambridge, Eng-
land) 141.1 (Jan. 2011), pp. 1–19. issn: 1741-7899. doi: 10.1530/REP-10-0236.
url: http://www.ncbi.nlm.nih.gov/pubmed/20974741.
[15] Gabor Vajta and Zsolt Peter Nagy. “Are programmable freezers still needed in the
embryo laboratory? Review on vitrification”. In: Reproductive BioMedicine Online
12.6 (Jan. 2006), pp. 779–796. issn: 14726483. doi: 10.1016/S1472-6483(10)610
91-7. url: http://linkinghub.elsevier.com/retrieve/pii/S1472648310610
917.
[16] Gabor Vajta and Masashige Kuwayama. “Improving cryopreservation systems.” In:
Theriogenology 65.1 (Jan. 2006), pp. 236–44. issn: 0093-691X. doi: 10.1016/j.t
heriogenology.2005.09.026. url: http://www.ncbi.nlm.nih.gov/pubmed/16
289262.
[17] W.F. Rall and G.M. Fahy. “Ice-free cryopreservation of mouse embryos at -196
degrees C by vitrification”. In: Nature 313.6003 (1985), pp. 573–5.
[18] Faten F AbdelHafez et al. “Slow freezing, vitrification and ultra-rapid freezing of
human embryos: a systematic review and meta-analysis.” In: Reproductive biomedicine
online 20.2 (Feb. 2010), pp. 209–22. issn: 1472-6491. doi: 10.1016/j.rbmo.200
9.11.013. url: http://www.ncbi.nlm.nih.gov/pubmed/20113959.
BIBLIOGRAPHY 76
[19] S Yavin and a Arav. “Measurement of essential physical properties of vitrification
solutions.” In: Theriogenology 67.1 (Jan. 2007), pp. 81–9. issn: 0093-691X. doi:
10.1016/j.theriogenology.2006.09.029. url: http://www.ncbi.nlm.nih.go
v/pubmed/17070573.
[20] George M Whitesides. “The origins and the future of microfluidics.” In: Nature
442.7101 (July 2006), pp. 368–73. issn: 1476-4687. doi: 10.1038/nature05058.
url: http://www.ncbi.nlm.nih.gov/pubmed/16871203.
[21] Curtis D Chin, Vincent Linder, and Samuel K Sia. “Commercialization of mi-
crofluidic point-of-care diagnostic devices.” In: Lab on a chip 12.12 (June 2012),
pp. 2118–34. issn: 1473-0189. doi: 10.1039/c2lc21204h. url: http://www.ncb
i.nlm.nih.gov/pubmed/22344520.
[22] Noha a Mousa et al. “Droplet-scale estrogen assays in breast tissue, blood, and
serum.” In: Science translational medicine 1.1 (Oct. 2009), 1ra2. issn: 1946-6242.
doi: 10.1126/scitranslmed.3000105. url: http://www.ncbi.nlm.nih.gov/pu
bmed/20368154.
[23] Mais J Jebrail and Aaron R Wheeler. “Let’s get digital: digitizing chemical biol-
ogy with microfluidics.” In: Current opinion in chemical biology 14.5 (Oct. 2010),
pp. 574–81. issn: 1879-0402. doi: 10.1016/j.cbpa.2010.06.187. url: http:
//www.ncbi.nlm.nih.gov/pubmed/20674472.
[24] Hanyoup Kim et al. “Automated digital microfluidic sample preparation for next-
generation DNA sequencing.” In: Journal of Laboratory Automation 16.6 (Dec.
2011), pp. 405–14. issn: 2211-0682. doi: 10.1016/j.jala.2011.07.001. url:
http://www.ncbi.nlm.nih.gov/pubmed/22093297.
[25] R. B. Fair. “Digital microfluidics: is a true lab-on-a-chip possible?” In: Microfluidics
and Nanofluidics 3.3 (Mar. 2007), pp. 245–281. issn: 1613-4982. doi: 10.1007/s1
BIBLIOGRAPHY 77
0404-007-0161-8. url: http://www.springerlink.com/index/10.1007/s1040
4-007-0161-8.
[26] SW Walker and Benjamin Shapiro. “Modeling the fluid dynamics of electrowetting
on dielectric (EWOD)”. In: Journal of Microelectromechanical Systems 15.4 (2006),
pp. 986–1000. url: http://ieeexplore.ieee.org/xpls/abs\_all.jsp?arnumb
er=1668195.
[27] Irena Barbulovic-Nad, Sam H Au, and Aaron R Wheeler. “A microfluidic plat-
form for complete mammalian cell culture.” In: Lab on a Chip 10.12 (June 2010),
pp. 1536–42. issn: 1473-0197. doi: 10.1039/c002147d. url: http://www.ncbi.n
lm.nih.gov/pubmed/20393662.
[28] Yi-Hsien Chang et al. “Integrated polymerase chain reaction chips utilizing digital
microfluidics.” In: Biomedical Microdevices 8.3 (Sept. 2006), pp. 215–25. issn: 1387-
2176. doi: 10.1007/s10544-006-8171-y. url: http://www.ncbi.nlm.nih.gov/
pubmed/16718406.
[29] Ramakrishna S Sista et al. “Heterogeneous immunoassays using magnetic beads on
a digital microfluidic platform.” In: Lab on a Chip 8.12 (Dec. 2008), pp. 2188–96.
issn: 1473-0197. doi: 10.1039/b807855f. url: http://www.pubmedcentral.ni
h.gov/articlerender.fcgi?artid=2726047\&tool=pmcentrez\&rendertype=a
bstract.
[30] Yun Seok Heo et al. “Controlled loading of cryoprotectants (CPAs) to oocyte with
linear and complex CPA profiles on a microfluidic platform.” In: Lab on a chip
11.20 (Sept. 2011), pp. 3530–7. issn: 1473-0189. doi: 10.1039/c1lc20377k. url:
http://www.ncbi.nlm.nih.gov/pubmed/21887438.
[31] D. Lai et al. “Automated microfluidic gradient cryoprotectant exchange platform
for murine oocyte and zygote vitrification reduces osmotic stress and improves
embryo developmental competence”. In: Fertility and Sterility 100.3 (Sept. 2013),
BIBLIOGRAPHY 78
S107. issn: 00150282. doi: 10.1016/j.fertnstert.2013.07.1682. url: http:
//linkinghub.elsevier.com/retrieve/pii/S0015028213024618.
[32] YS Song et al. “Microfluidics for cryopreservation”. In: Lab on a Chip 9.13 (2009),
pp. 1874–1881. doi: 10.1039/b823062e.Microfluidics. url: http://pubs.rs
c.org/en/content/articlehtml/2009/lc/b823062e.
[33] Liu Chang. Foundation of MEMS. 2nd. Pearson Education Limited, 2011. isbn:
0132497360.
[34] J Ali and J N Shelton. “Design of vitrification solutions for the cryopreservation
of embryos.” In: Journal of reproduction and fertility 99.2 (Nov. 1993), pp. 471–7.
issn: 0022-4251. url: http://www.ncbi.nlm.nih.gov/pubmed/8107029.
[35] J Berthier, Ph Clementz, and J-m Roux. “Modeling microdrop motion between
covered and open regions of EWOD microsystems”. In: NSTI Nanotechnology Con-
ference and Trade Show-Nanotech, NSTI: Boston, USA. Vol. 1. 2006, pp. 685–688.
isbn: 0976798565. url: http://nrc.org/publications/Nanotech/2006/pdf/36
0.pdf.
[36] J E Swain et al. “Thinking big by thinking small: application of microfluidic technol-
ogy to improve ART.” In: Lab on a chip 13.7 (Mar. 2013), pp. 1213–24. issn: 1473-
0189. doi: 10.1039/c3lc41290c. url: http://www.ncbi.nlm.nih.gov/pubmed/
23400523.
[37] Brigid Hogan et al. Manipulating the mouse embryo: a laboratory manual. 3rd.
Vol. 34. Cold Spring Harbor, NY: Cold spring harbor laboratory, 1986, p. 800.
isbn: 0879695919.
[38] A Dhali et al. “Effect of droplet vitrification on development competence , actin
cytoskeletal integrity and gene expression in in vitro cultured mouse embryos”. In:
Theriogenology 71 (2009), pp. 1408–1416. doi: 10.1016/j.theriogenology.200
9.01.011.
BIBLIOGRAPHY 79
[39] Junqiang Zhang et al. “Vitrification of mouse embryos at 2-cell, 4-cell and 8-cell
stages by cryotop method.” In: Journal of assisted reproduction and genetics 26.11-
12 (2009), pp. 621–8. issn: 1573-7330. doi: 10.1007/s10815-009-9370-2. url:
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2799565\
&tool=pmcentrez\&rendertype=abstract.
[40] T Otoi et al. “Cryopreservation of mature bovine oocytes by vitrification in straws.”
In: Cryobiology 37.1 (Aug. 1998), pp. 77–85. issn: 0011-2240. doi: 10.1006/cryo.
1998.2103. url: http://www.ncbi.nlm.nih.gov/pubmed/9698432.
[41] K Nakao, N Nakagata, and M Katsuki. “Simple and efficient vitrification procedure
for cryopreservation of mouse embryos”. In: Experimental animals (1997). url:
http://japanlinkcenter.org/JST.JSTAGE/expanim/46.231?from=Google\
&type=all\_cited\&lang=en.
[42] Masashige Kuwayama et al. “Comparison of open and closed methods for vitri-
fication of human embryos and the elimination of potential contamination”. In:
Reproductive BioMedicine Online 11.5 (Jan. 2005), pp. 608–614. issn: 14726483.
doi: 10.1016/S1472-6483(10)61169-8. url: http://linkinghub.elsevier.co
m/retrieve/pii/S1472648310611698.
[43] a Martino, N Songsasen, and S P Leibo. “Development into blastocysts of bovine
oocytes cryopreserved by ultra-rapid cooling.” In: Biology of reproduction 54.5 (May
1996), pp. 1059–69. issn: 0006-3363. url: http://www.ncbi.nlm.nih.gov/pubme
d/8722627.
[44] Xiaoming Zhou et al. “Theoretical investigations of a novel microfluidic cooling/warming
system for cell vitrification cryopreservation”. In: International Journal of Heat
and Mass Transfer 65 (Oct. 2013), pp. 381–388. issn: 00179310. doi: 10 . 101
6/j.ijheatmasstransfer.2013.06.022. url: http://linkinghub.elsevier.c
om/retrieve/pii/S0017931013004900.
BIBLIOGRAPHY 80
[45] Young S Song et al. “Vitrification and levitation of a liquid droplet on liquid nitro-
gen.” In: Proceedings of the National Academy of Sciences of the United States of
America 107.10 (Mar. 2010), pp. 4596–600. issn: 1091-6490. doi: 10.1073/pnas.
0914059107. url: http://www.pubmedcentral.nih.gov/articlerender.fcgi?
artid=2826340\&tool=pmcentrez\&rendertype=abstract.
[46] Vitrolife. Rapid-i Vitrification of Cleavage stage embryos. 2011. url: http://ww
w.vitrolife.com/en/Fertility/Procedures/Vitrification/.
[47] Canadian Mouse Mutant Repository. Modified Rapid Embryo Freezing Protocol
D for ENU and Non-Enu Mice Embryos at Eight-Cell Stage. 2004. url: http:
//www.cmmr.ca/protocols/embryo\_cryo\_D.pdf.
[48] Somayeh Khosravi-Farsani et al. “Mouse oocyte vitrification: the effects of two
methods on maturing germinal vesicle breakdown oocytes.” In: Journal of assisted
reproduction and genetics 27.5 (May 2010), pp. 233–8. issn: 1573-7330. doi: 10.1
007/s10815-010-9411-x. url: http://www.pubmedcentral.nih.gov/articler
ender.fcgi?artid=2881200\&tool=pmcentrez\&rendertype=abstract.
[49] Irvine Scientific. Simplified Embryo Vitrification Protocol. 2011. url: http://w
ww.irvinesci.com/products/90133- dso- vitrification- freeze- solutio
ns- for- embryo- pn- to- blastocyst- stage?dpt=Assisted+Reproductive+
Technology.
[50] Origio. MediCult Vitrification Cooling. 2011. url: http://www.origio.com/prod
ucts/medicultmedia/medicultvitrificationcooling.aspx.
[51] Kitazato Corp. Cryotop Safety Kit. 2012. url: http://www.kitazato.co.jp/bio
pharma/VT60.html.
[52] Masashige Kuwayama. “Highly efficient vitrification for cryopreservation of human
oocytes and embryos: the Cryotop method.” In: Theriogenology 67.1 (Jan. 2007),
BIBLIOGRAPHY 81
pp. 73–80. issn: 0093-691X. doi: 10.1016/j.theriogenology.2006.09.014. url:
http://www.ncbi.nlm.nih.gov/pubmed/17055564.
[53] Sam H Au, Paresh Kumar, and Aaron R Wheeler. “A new angle on pluronic addi-
tives: advancing droplets and understanding in digital microfluidics.” In: Langmuir :
the ACS journal of surfaces and colloids 27.13 (July 2011), pp. 8586–94. issn: 1520-
5827. doi: 10.1021/la201185c. url: http://www.ncbi.nlm.nih.gov/pubmed/2
1651299.
[54] Vivienne N Luk, Gary Ch Mo, and Aaron R Wheeler. “Pluronic additives: a solution
to sticky problems in digital microfluidics.” In: Langmuir : the ACS journal of
surfaces and colloids 24.12 (June 2008), pp. 6382–9. issn: 0743-7463. doi: 10.102
1/la7039509. url: http://www.ncbi.nlm.nih.gov/pubmed/18481875.
[55] Vijay Srinivasan, Vamsee K Pamula, and Richard B Fair. “An integrated digital
microfluidic lab-on-a-chip for clinical diagnostics on human physiological fluids.” In:
Lab on a chip 4.4 (Aug. 2004), pp. 310–5. issn: 1473-0197. doi: 10.1039/b403341h.
url: http://www.ncbi.nlm.nih.gov/pubmed/15269796.
[56] M Jonsson-Niedzioka et al. “EWOD driven cleaning of bioparticles on hydrophobic
and superhydrophobic surfaces.” In: Lab on a chip 11.3 (Feb. 2011), pp. 490–6.
issn: 1473-0189. doi: 10.1039/c0lc00203h. url: http://www.ncbi.nlm.nih.go
v/pubmed/21103534.
[57] Andrea E Kirby and Aaron R Wheeler. “Microfluidic origami: a new device format
for in-line reaction monitoring by nanoelectrospray ionization mass spectrometry.”
In: Lab on a chip 13.13 (July 2013), pp. 2533–40. issn: 1473-0189. doi: 10.1039/c
3lc41431k. url: http://www.ncbi.nlm.nih.gov/pubmed/23412052.
[58] Mohamed Abdelgawad et al. “All-terrain droplet actuation.” In: Lab on a chip 8.5
(May 2008), pp. 672–7. issn: 1473-0197. doi: 10.1039/b801516c. url: http:
//www.ncbi.nlm.nih.gov/pubmed/18432335.
BIBLIOGRAPHY 82
[59] M G Pollack, a D Shenderov, and R B Fair. “Electrowetting-based actuation of
droplets for integrated microfluidics.” In: Lab on a chip 2.2 (May 2002), pp. 96–
101. issn: 1473-0197. doi: 10.1039/b110474h. url: http://www.ncbi.nlm.nih.
gov/pubmed/15100841.
[60] AR Wheeler et al. “Digital microfluidics with in-line sample purification for pro-
teomics analyses with MALDI-MS”. In: Analytical chemistry 77.2 (2005), pp. 534–
540. url: http://pubs.acs.org/doi/abs/10.1021/ac048754\%2B.
[61] Sam H Au et al. “Cellular bias on the microscale: probing the effects of digital
microfluidic actuation on mammalian cell health, fitness and phenotype.” In: Inte-
grative biology 5.8 (Aug. 2013), pp. 1014–25. issn: 1757-9708. doi: 10.1039/c3ib
40104a. url: http://www.ncbi.nlm.nih.gov/pubmed/23770992.
[62] Hyejin Moon et al. “Low voltage electrowetting-on-dielectric”. In: Journal of Ap-
plied Physics 92.7 (2002), p. 4080. issn: 00218979. doi: 10.1063/1.1504171. url:
http://link.aip.org/link/JAPIAU/v92/i7/p4080/s1\&Agg=doi.
[63] Sung Kwon Cho, Hyejin Moon, and Chang-jin Kim. “Creating, transporting, cut-
ting, and merging liquid droplets by electrowetting-based actuation for digital mi-
crofluidic circuits”. In: Journal of Microelectromechanical Systems 12.1 (Feb. 2003),
pp. 70–80. issn: 1057-7157. doi: 10 . 1109 / JMEMS . 2002 . 807467. url: http :
//ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1183744.
[64] Michael G Pollack, Richard B Fair, and North Carolina. “Electrowetting-based ac-
tuation of liquid droplets for microfluidic applications”. In: Applied Physics Letters
77.11 (2000), pp. 1725–1726.
[65] Mohamed Abdelgawad, Philip Park, and Aaron R. Wheeler. “Optimization of de-
vice geometry in single-plate digital microfluidics”. In: Journal of Applied Physics
105.9 (2009), p. 094506. issn: 00218979. doi: 10.1063/1.3117216. url: http:
//link.aip.org/link/JAPIAU/v105/i9/p094506/s1\&Agg=doi.
BIBLIOGRAPHY 83
[66] Thomas Lederer et al. “Integration of impedance spectroscopy sensors in a digital
microfluidic platform”. In: Microsystem Technologies 18.7-8 (Mar. 2012), pp. 1163–
1180. issn: 0946-7076. doi: 10.1007/s00542-012-1464-6. url: http://link.sp
ringer.com/10.1007/s00542-012-1464-6.
[67] Sung Kwon Cho, Yuejun Zhao, and C.J. Kim. “Concentration and binary separa-
tion of micro particles for droplet-based digital microfluidics.” In: Lab on a Chip
7.4 (Apr. 2007), pp. 490–8. issn: 1473-0197. doi: 10.1039/b615665g. url: http:
//www.ncbi.nlm.nih.gov/pubmed/17389966.
[68] Huijiang Ding et al. “Accurate dispensing of volatile reagents on demand for chem-
ical reactions in EWOD chips.” In: Lab on a Chip 12.18 (Sept. 2012), pp. 3331–40.
issn: 1473-0189. doi: 10.1039/c2lc40244k. url: http://www.ncbi.nlm.nih.go
v/pubmed/22825699.
[69] Kyungyong Choi et al. “Integration of field effect transistor-based biosensors with
a digital microfluidic device for a lab-on-a-chip application.” In: Lab on a Chip
12.8 (Apr. 2012), pp. 1533–9. issn: 1473-0189. doi: 10.1039/c2lc21203j. url:
http://www.ncbi.nlm.nih.gov/pubmed/22402581.
[70] David J. Griffiths. Introduction to Electrodynamics. 3rd. Upper Saddle River, N.J.:
Prentice Hall, 1999.
[71] Patrick M. Young and Kamran Mohseni. “Calculation of DEP and EWOD Forces
for Application in Digital Microfluidics”. In: Journal of Fluids Engineering 130.8
(2008), p. 081603. issn: 00982202. doi: 10.1115/1.2956606. url: http://link.
aip.org/link/JFEGA4/v130/i8/p081603/s1\&Agg=doi.
[72] Mohamed Abdelgawad and Aaron R. Wheeler. “The Digital Revolution: A New
Paradigm for Microfluidics”. In: Advanced Materials 21.8 (Feb. 2009), pp. 920–925.
issn: 09359648. doi: 10.1002/adma.200802244. url: http://doi.wiley.com/1
0.1002/adma.200802244.