FINGER-POWERED THERMOPLASTIC MICROFLUIDIC
ELECTROCHEMICAL ASSAY FOR DIAGNOSTIC TESTING
USING A MOBILE PHONE
By
Tung-Yi Lin

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Mechanical Engineering – Doctor of Philosophy
2018

ABSTRACT
FINGER-POWERED THERMOPLASTIC MICROFLUIDIC ELECTROCHEMICAL ASSAY
FOR DIAGNOSTIC TESTING USING A MOBILE PHONE
By
Tung-Yi Lin
Point-of-care (POC) testing has gained considerable attention in recent years due to its
ability to provide diagnostic information without the need for centralized laboratory facilities or
bulky equipment. This has been achieved, in part, by advances in micro-electro-mechanical system
(MEMS) and analytical chemistry, which has resulted in the miniaturization and integration of
sensitive biosensors and fluidic components. Recently, researchers have demonstrated the use of
mobile phones for POC testing, which offers the advantages of portability and wireless data
transmission. Many mobile phone-based POC tests are based on optical imaging or colorimetric
assays, which are useful for some diagnostic applications, but lack the accuracy and sensitivity
required for the diagnosis of many important diseases. Moreover, these devices employ
microfluidic chips fabricated using glass, polydimethylsiloxane (PDMS) or paper, which require
complex microfabrication or surface treatments, or offer limited fluidic control.
In this dissertation, we explored the development of plastic-based microfluidic chips for
rapid electrochemical measurements of protein biomarkers using a mobile phone biosensing
platform. We first investigated UV/ozone (UVO) surface treatment on plastics to better understand
its usefulness for microfluidic POC applications. We found that UVO-treated poly(methyl
methacrylate) (PMMA), cyclic olefin copolymer (COC) and polycarbonate (PC) experience
hydrophobic recovery within 4 weeks and the rate at which it occurs is dependent on the UVO
treatment duration. Furthermore, we discovered that the hydrophobic recovery of UVO-treated
COC and PC can be inhibited by storing them in dehumidified or vacuum conditions. UVO-treated
plastics were also used for protein adsorption measurements, which showed that UVO treatment
minimized protein adsorption and this effect is correlated with the treatment duration. Lastly, we

demonstrated capillary-driven flows in UVO-treated PMMA microchannels, which revealed that
the flow rate can be tuned by adjusting the treatment duration.
We also explored the development of new fabrication methods for generating plastic
microfluidic devices. In particular, we have demonstrated for the first time the use of 3D printed
metal molds for fabricating plastic microchannels via hot embossing. Through the optimization of
the powder composition and processing parameters, we generated stainless steel molds with
superior material properties (density and surface finish) and replication accuracy compared with
previously reported 3D printed metal parts. 3D printed molds were used to fabricate PMMA
replicas, which exhibited good feature integrity and replication quality. Microchannels fabricated
using these replicas exhibited leak-free operation and comparable flow performance as
microchannels fabricated from CNC milled molds for both capillary and pressure-driven flows.
Toward the realization of a shelf stable, electricity-free microfluidic assay for POC testing,
we developed a finger-powered microfluidic chip for electrochemical measurements of protein
biomarkers. This device employs a valveless, piston-based pumping mechanism which utilizes a
human finger for the actuation force. Liquids are driven inside microchannels by pressing on a
mechanical piston, which generates a pressure-driven flow. Dried reagents are preloaded in
microwells allowing for the entire testing process to be completed on-chip. Additionally, a
nonenzymatic detection scheme is employed which circumvents the need for refrigeration. For
proof-of-concept, this microfluidic assay was coupled with a mobile phone biosensing platform for
quantitative measurements of Plasmodium falciparum histidine-rich protein-2 (PfHRP2) in human
blood samples. Using this platform, PfHRP2 was detected from 1 to 20 Âľg/mL with high specificity
and each measurement could be completed in ≤ 5 min. In addition, this assay can be stored at room
temperature for up to one month with a negligible loss in performance. The results and knowledge
presented in this dissertation will provide new insights into the development of plastic microfluidic
devices for POC testing as well as other biomedical application.

ACKNOWLEDGEMENTS

This dissertation is the product of my four years efforts since I joined the MSU program in
2013. It was a truly great memory in my life. First and foremost, I would like to thank the guidance
of my advisor, Dr. Peter Lillehoj. He has taught me a lot regarding to the projects, course works
and the career paths. He is knowledgeable and willing to support all the students. I’m really grateful
for his contributions through my PhD years. In addition, I would like to thank all my committee
members, including Dr. Wen Li, Dr. Junghoon Yeom and Dr. Manoochehr Koochesfahani. Without
they’re insightful inputs and valuable comments to my works, I couldn’t complete this project.
My gratitude also goes to my labmates that I have met in MSU. First, I would like to Xiyuan
Liu, my first labmate, who always generously shared her research experience with me. I also
appreciate her advise and help through my PhD years. Furthermore, thanks to Zoe Jiang for her
efforts and helps for maintaining the lab. I would like to thank to Trey Pfeiffer, an undergraduate
student in mechanical engineering department. He’s the smartest undergraduate that I have met in
MSU. In addition to his tremendous contribution in the projects, he’s actually one of my best friends.
He is willing to share his life and entertainments with me and make my PhD career interesting.
Moreover, I would like to extend my acknowledgements to all the faculties supported me
within four years. Thanks to one of my collaborators, Dr. Patrick Kwon provided suggestions and
supports to the 3D metal printing project. Thanks to Kathryn Severin for her assistance with UVvisible spectroscopy and contact angle measurements. Thanks to Dr. Baokang Bi and Matthew
Oonk for their assistance with cleanroom facilities. Thanks to Brian Wright for his supports with
the 3D printing and surface profilometer measurement. Thanks to Dr. Per Askeland for his
assistance with XPS measurement and element analysis. Thanks to Dr. Reza Loloee and Dr.
Melinda Frame for the assistance of microscopy.

iv

Thanks to my main funding source National Institutes of Health organization. This work
was funded by the NIH (R01AI113257). I would like to thank to my teammates involved in this
project, Sina Parsenjad and Linlin Tu. Sina Parsenjad helped me with development of couple
generation circuit devices. He is a friendly person and reachable all the time when I faced electrical
problems. Linlin developed a phone application and constantly improved the application interface.
For development the prototype, we had many difficulties, and finally achieved all these challenges.
I sincerely thank to their patience and contributions for this project. Moreover, I thank Dr.
Gorachand Dutta for his assistance with designing and troubleshooting the detection scheme and
also thank Dr. Hao Wan for developing the PC software used for recording the sensor data. Thanks
to all the collaborators involved in this project, including Dr. Andrew Mason, Dr. Guoliang Xing,
Dr. Terrie Taylor, Dr. Karl Seydel and Dr. Mat Reeves.
Lastly, I would like to thank to my family, my friends and my girlfriend, Ting-Wei Lee. I
really appreciate their unconditional love and encouragement. They accompany me through entire
my PhD career, and bring me a powerful strength in my life.

v

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES ........................................................................................................................ ix
Chapter 1. Introduction.................................................................................................................... 1
1.1 Point-of-care testing .............................................................................................................. 1
1.2 Electrochemical biosensors ................................................................................................... 3
1.2.1 Immobilization of bioreceptors ...................................................................................... 4
1.2.2 Principle of electrochemical sensing .............................................................................. 5
1.2.3 Amperometric biosensors ............................................................................................... 6
1.3 Microfluidic-based electrochemical biosensors .................................................................... 7
1.4 Electricity-free microfluidic pumping ................................................................................... 9
1.4.1 Capillary flow ................................................................................................................. 9
1.4.2 Finger-powered pumping ............................................................................................... 9
1.5 Mobile phone-based biosensors .......................................................................................... 10
1.6 Objectives ............................................................................................................................ 13
Chapter 2. Investigation of UVO treatment on thermoplastics ..................................................... 15
2.1 Motivation ........................................................................................................................... 15
2.2 Methods and materials......................................................................................................... 15
2.2.1 UVO treatment and sample storage .............................................................................. 15
2.2.2 Contact angle measurements ........................................................................................ 16
2.2.3 XPS analysis and AFM measurements ......................................................................... 16
2.2.4 Protein adsorption measurements ................................................................................. 16
2.3 Surface chemistry and morphology of UVO-treated plastics .............................................. 17
2.4 Wetting stability of UVO-treated surface............................................................................ 20
2.5 Effectiveness of UVO treatment for capillary flow and protein retention .......................... 23
2.6 Summary.............................................................................................................................. 26
Chapter 3. Plastic-based microfluidic device fabrication methods ............................................... 27
3.1 Motivation ........................................................................................................................... 27
3.2 Hot embossing using 3D printed molds with thermal bonding ........................................... 28
3.2.1 Materials and methods .................................................................................................. 28
3.2.2 Fabrication and characterization of 3D printed molds ................................................. 30
3.2.3 Characterization of PMMA replicas ............................................................................. 31
3.2.4 Thermal bonding and the influence of surface roughness ............................................ 32
3.2.5 Microfluidic device testing ........................................................................................... 34
3.3 Laser ablation with adhesive lamination bonding ............................................................... 36
3.3.1 Materials and methods .................................................................................................. 36
3.3.2 Optimization of operating settings for laser ablation of PMMA .................................. 37
3.4 Summary.............................................................................................................................. 39
Chapter 4. Electrochemical assay development for PfHRP2 quantification ................................. 40
4.1 Motivation ........................................................................................................................... 40
4.2 Horseradish peroxidase-based electrochemical assay ......................................................... 41
4.2.1 Materials and methods .................................................................................................. 41
4.2.2 Antibody immobilization optimization ........................................................................ 42
vi

4.2.3 Substrate optimization .................................................................................................. 50
4.3 Methylene blue-based electrochemical assay ...................................................................... 53
4.3.1 Materials and methods .................................................................................................. 54
4.3.2 Antibody immobilization optimization ........................................................................ 55
4.3.3 Substrate optimization .................................................................................................. 55
4.3.4 Electrical field preparation optimization ...................................................................... 56
4.4 Summary.............................................................................................................................. 59
Chapter 5. Finger-powered microfluidic pump ............................................................................. 61
5.1 Motivation ........................................................................................................................... 61
5.2 Materials and methods......................................................................................................... 61
5.2.1 Finger-powered pump fabrication ................................................................................ 61
5.2.2 Finger-powered pump characterization ........................................................................ 61
5.3 PDMS membrane-based pump ............................................................................................ 62
5.4 Piston-based pump .............................................................................................................. 64
5.5 Gasket-based piston pump ................................................................................................... 66
5.6 Summary.............................................................................................................................. 73
Chapter 6. Microfluidic cartridge development ............................................................................ 74
6.1 Motivation ........................................................................................................................... 74
6.2 Materials and methods......................................................................................................... 74
6.2.1 Microfluidic cartridge design ....................................................................................... 74
6.2.2 Microfluidic cartridge fabrication ................................................................................ 74
6.2.3 Amperometric measurements ....................................................................................... 75
6.3 First generation prototype device ........................................................................................ 76
6.3.1 Cartridge design............................................................................................................ 76
6.3.2 Device testing ............................................................................................................... 78
6.3.3 Summary ...................................................................................................................... 80
6.4 Second generation prototype device .................................................................................... 81
6.4.1 Cartridge design............................................................................................................ 81
6.4.2 Device testing ............................................................................................................... 84
6.4.3 Device validation .......................................................................................................... 87
6.5 Summary.............................................................................................................................. 88
Chapter 7. Conclusions.................................................................................................................. 90
7.1 Summary of achievements and contributions ...................................................................... 90
7.2 Future work ......................................................................................................................... 91
BIBLIOGRAPHY ......................................................................................................................... 92

vii

LIST OF TABLES

Table 2-1. XPS analysis of the O/C ratio on the surface of UVO-treated and untreated plastics
immediately following UVO exposure. ........................................................................................ 19
Table 2-2. Surface roughness values (nm) of UVO-treated and untreated plastic samples
immediately after treatment and after 16 weeks of storage under different storage conditions. ... 20
Table 2-3. XPS analysis of the O/C ratio on the surface of UVO-treated and untreated plastics
after 16 weeks under different storage conditions. ........................................................................ 23
Table 3-1. Ablation feature size with respect to different power settings on PMMA. Each
measurement is the average of five different samples. .................................................................. 38
Table 3-2. Ablation feature size with respect to number of passes 42W on PMMA. Each
measurement is the average of five different samples. .................................................................. 38

viii

LIST OF FIGURES

Figure 1-1. Procedures for conventional lab-based diagnostics compared to point-of-care (POC)
testing. Image obtained from [1]. .................................................................................................... 1
Figure 1-2. Major elements of a typical biosensor. Image obtained from [4]. ................................ 2
Figure 1-3. Process flow of a typical biosensor. Image obtained from [3]. .................................... 4
Figure 1-4. Schematic of main methods for immobilizing bioreceptors on a senor surface. Image
obtained from [11]. .......................................................................................................................... 4
Figure 1-5. Example of an electrochemical measurement process for the detection of lactate.
Image obtained from [9]. ................................................................................................................. 5
Figure 1-6. Schematic of three types of finger-powered pumping mechanisms: (a) screw-based
pumping, image obtained from [39], (b) vacuum activation pumping, image obtained from [40],
and (c) push and release pumping, image obtained from [41]. ..................................................... 10
Figure 1-7. Examples of mobile phone-based systems for POC protein measurements. (a)
Fluorescent measurement for quantifying albumin in urine. Image obtained from [47]. (b) Opticscatter detection system for measuring protein amount. Image obtained from [48]. (c)
Electrochemical measurement for evaluating malaria protein in human plasma. Image obtained
from [49]. (d) Optical measurement on electrode for quantifying HIV protein. Image obtained
from [41]........................................................................................................................................ 13
Figure 2-1. Contact angles of PMMA, COC and PC with varying durations of UVO treatment.
Each bar represents the mean Âą standard deviation (SD) of eight separate measurements. .......... 18
Figure 2-2. AFM tapping mode images of untreated PMMA (a), PC (b), and COC (c), and UVOtreated (80 min) PMMA (d), PC (e) and COC (f). The scan size and z-scale are 4 µm × 4 µm and
50 nm, respectively........................................................................................................................ 19
Figure 2-3. Long-term contact angle measurements of PMMA, COC and PC treated for 20 min
(a), 40 min (b), 60 min (c) and 80 min (d) under various storage conditions. Black, grey and white
markers correspond to storage in air, dehumidified and vacuum conditions, respectively. Each
data point represents the mean Âą SD of eight separate measurements. ......................................... 22
Figure 2-4. Capillary flow in untreated (left) and 80 min UVO-treated (right) PMMA
microchannels. Colored dye is dispensed at the inlet of the chips using a pipette. ....................... 23
Figure 2-5. Capillary flow rate as a function of UVO treatment duration for PMMA
microchannels. Inset shows magnified view of the data points at lower (< 40 min) UVO exposure
times. Each data point represents the mean Âą SD of three measurements. .................................. 24
Figure 2-6. Absorbance values of TMB/H2O2 on plastics exposed to UVO at varying durations.
Each data point represents the mean Âą SD of three separate measurements obtained using new
samples. ......................................................................................................................................... 25

ix

Figure 3-1. Photographs (upper) and surface profiles (lower) of 3D printed molds containing 0%
(a), 0.25% (b), 0.5% (c) and 0.75% (d) wt. BN. The scan size and z-scale are 600 µm × 600 µm
and 120 Âľm respectively................................................................................................................ 31
Figure 3-2. Photographs (upper) and surface profiles (lower) of PMMA replicas fabricated from
3D printed molds containing 0% (a), 0.25% (b), 0.5% (c) and 0.75% (d) wt. BN. The scan size
and z-scale are 600 µm × 600 µm and 120 µm, respectively. ....................................................... 32
Figure 3-3. Close up images of embossed microfluidic features in PMMA fabricated using the
pure stainless steel mold (a) and mold containing 0.25% wt. BN (b). Hot embossing was
performed at 120°C and 1 ton for 30 min. Scale bars, 500 ¾m. ................................................... 32
Figure 3-4. SEM images of cross sections of PMMA microchannels fabricated using the pure
stainless steel mold and thermally bonded at 80°C (a), 90°C (b) and the mold containing 0.25%
wt. BN at 80°C (c). The arrow indicates incomplete bonding. Scale bars, 500 ¾m. ..................... 33
Figure 3-5. (a) Capillary flow in PMMA microchannels fabricated from 3D printed molds.
Colored dye is dispensed at the inlet of the chips using a pipette. (b) Pressure-driven flow inside a
PMMA microchannel fabricated from a 3D printed mold. ........................................................... 34
Figure 3-6. Comparison of flow performance in PMMA microchannels fabricated from 3D
printed molds (squares) and CNC milled molds (circles). Capillary flow rates and pressure-driven
flow rates are plotted as hollow and solid markers, respectively. Each data point represents the
mean Âą SD of three measurements for capillary flow experiments and five measurements for
pressure-driven flow experiments. ................................................................................................ 35
Figure 3-7. Microfluidic cartridge fabrication process based on CO2 laser ablation and adhesive
lamination bonding. ....................................................................................................................... 37
Figure 4-1. Schematic of sandwich ELISA. Target protein is encapsulated between two layers
antibodies, followed by TMB reacting with Px and generate signal. The signal should be
proportional to the concentration of target protein. Image obtained from [88]. ............................ 40
Figure 4-2. Schematic of an enzymatic PfHRP2 electrochemical assay. (a) The sensor is
immobilized with capture antibody on surface. (b) If present, sample binds to the capture
antibody. (c) Detection antibody is immobilized on the sensor via binding to sample, forming a
sandwich structure. (d) Peroxidase reacts with substrate generating a current that is proportional
to the PfHRP2 concentration. ........................................................................................................ 41
Figure 4-3. Different methods for performing immobilization and functional groups used for
orienting antibodies on substrate [62] ........................................................................................... 43
Figure 4-4. Amperometric signals of PfHRP2 diluted in PBS using sensors prepared by physical
absorption immobilization technique. Each bar represents the mean Âą SD of four measurements.
....................................................................................................................................................... 44
Figure 4-5. Schematic of 11-MUA self-assembly on gold electrode and subsequent
immobilization of capture antibody and antigen. Image obtained from [93]. ............................... 45
Figure 4-6. Amperometric signals of PfHRP2 diluted in PBS via amine- group covalent binding
immobilization technique. Each bar represents the mean Âą SD of four measurements. ............... 46
x

Figure 4-7. Amperometric signals of PfHRP2 diluted in PBS obtained from electric field-assisted
sensors (solid) and sensors undergoing incubation. Each bar represents the mean Âą SD of two
measurements. ............................................................................................................................... 47
Figure 4-8. Amperometric signals of PfHRP2 diluted in PBS (black), human plasma (striped) and
human blood (gray). Each bar represents the mean Âą SD of four measurements.......................... 48
Figure 4-9. Schematic of the working mechanism for thiol group immobilization method. Image
modified from [103]. ..................................................................................................................... 49
Figure 4-10. Amperometric sensing of PfHRP2 diluted in human blood based on carboxyl group
antibody immobilization (solid) and thiol-group antibody immobilization (striped). Each bar
represents the mean Âą SD of two measurements. .......................................................................... 50
Figure 4-11. Schematic of the detection scheme for PfHRP2 detection based on a redox reaction
between an enzyme reporter and TMB+H2O2. Image obtained from [97]. ................................... 51
Figure 4-12. Amperometric results were obtained from spiking sample in PBS. Signal-tobackground ratio was used for comparing different substrate of 4AP and TMB. There is
negligible signal difference for TMB for concentration above 1 Âľg/mL. Each bar represents the
mean Âą SD of two measurements. ................................................................................................. 52
Figure 4-13. Amperometric signals of PfHRP2 diluted in human plasma using a bias potential of 200 mV (solid) and -150 mV (striped). The scheme is based on thiol immobilization method
using 4AP as the substrate. Each bar represents the mean Âą SD of two measurements. ............... 53
Figure 4-14. Schematic of electrochemical redox cycling based on the redox reaction between
MB and Ru(NH3)63+ . ..................................................................................................................... 54
Figure 4-15. Schematic of electron transfer scheme on different antigen–antibody binding
conditions. (a) When PfHRP2 is presented on surface, the electron transfer is inhibited. (b) When
PfHRP2 is absence, the electron transfer is higher. ....................................................................... 56
Figure 4-16. Effect of different duration of 400 mV driving potential on whole blood samples
containing 10 Îźg/mL for comparing the value of target and without target. Each bar represents the
mean Âą SD of three separate measurements obtained using new sensors. .................................... 57
Figure 4-17. Cyclic voltammograms of whole blood samples containing 10 Îźg/mL with electrical
field preparation (EFP) and without EFP using open well chips. ................................................. 58
Figure 4-18. Cyclic voltammograms of whole blood samples spiked with 10 Îźg/mL PfHRP2 and
without PfHRP2. The measurement is performed using a microfluidic cartridge......................... 59
Figure 5-1. Photograph of experimental setup for piston actuation force measurements. The
compression rod is aligned with the piston, and the resulting data is transmitted to a PC. The inset
shows a close-up view of the compression rod pressing the piston during a measurement. ......... 62
Figure 5-2. Photograph of the PDMS membrane-based pump before (left) and after(right) finger
actuation. ....................................................................................................................................... 63

xi

Figure 5-3. (a)(b) Schematic of the working principle of the fluidic valve to regulate flow. (c) Top
view the one-way valve captured by optical microscopy. Scale bar, 500 Îźm. .............................. 63
Figure 5-4. Microfluidic cartridge testing using color dyes. (a) The green dye was dispensed
toward outlet. The arrow indicates poor liquid coverage for sensor region. (b)(c) Red and green
were dispensed sequentially. The arrow reveals blue dye leaking resulting from incomplete
bonding. ......................................................................................................................................... 64
Figure 5-5. Schematics of the three different designs of piston-based finger-powered pumps. Red
circles indicates defects during actuation. (a) The PDMS on the outer layer breaks due to the
friction force. (b) The depressed piston is tilting and results in leaking. (c) The optimal pump
design............................................................................................................................................. 65
Figure 5-6. Microfluidic cartridge testing with colored dye. (a) Red dye was dispensed via
capillary tube, and blue dye (b) DI water (c) and yellow dye (d) were sequentially dispensed by
depressing the pistons. ................................................................................................................... 66
Figure 5-7. Schematic of the modified piston-based microfluidic pump before (left) and after
(right) actuation. Arrow indicates the rubber-based gasket to prevent liquid leaking. .................. 66
Figure 5-8. Experimental vs. theoretical flow rates for rectangular microchannels 100 Âľm in
height, 15 cm long with varying widths. Each data point is the mean Âą SD of five measurements
using new chips. ............................................................................................................................ 69
Figure 5-9. Flow rate vs. pressing force for rectangular microchannels 100 Âľm high, 400 Âľm wide
and 10 cm long. Each bar represents the mean Âą SD of five separate measurements. ................. 70
Figure 5-10. Pressing force required to pump water in rectangular microchannels 100 Âľm high, 10
cm long and of varying widths. Each data point represents the mean Âą SD of three separate
measurements. ............................................................................................................................... 71
Figure 5-11. (a) Schematic of microfluidic device for evaluating washing efficiency. Fluorescent
images of the reservoir before (b) and after (c) washing. Scale bars, 100 Âľm. (d) Washing
efficiency in rectangular microchannels at 100 Âľm high and 400 Âľm of varying weights. Each bar
represents the mean Âą SD of three separate measurements. .......................................................... 72
Figure 6-1. Photograph of our mobile phone biosensing platform. The microfluidic chip (a)
directly plugs into aMEASURE (b), which is connected to the phone via the microUSB port using
a USB OTG adapter. ..................................................................................................................... 75
Figure 6-2. Exploded view of microfluidic cartridge (left). Photograph of the assembled
microfluidic cartridge with relevant regions labeled. Reservoirs R1, R2 and R3 are designed to
preload with a reporter solution, wash buffer and 4AP/H2O2 substrate, respectively. The
dimensions of the cartridge are 40 mm × 30 mm. ......................................................................... 77
Figure 6-3. Microfluidic cartridge testing with color dye. Red dye was added by capillary tube,
and blue dye, DI water and yellow dye were sequentially dispensed by depressing the pistons. (a)
The sample is dispensed from capillary tube and flows toward the sensor region. (b) Detection
antibody solution flows to the sensor by pressing piston R1. (c) Buffer washes away unbound
detection antibodies by pressing piston R2 (d) Substrate flows to the sensor region by pressing
piston R3........................................................................................................................................ 77
xii

Figure 6-4. Specificity of PfHRP2 assay. Measurements were performed using DEN-2, PfLDH
and PfHRP2 in PBS at a concentration of 10 Îźg/mL, and PBS as a blank control. Each bar
represents of the mean Âą SD of three separate measurements. ...................................................... 79
Figure 6-5. Amperometric signals of PfHRP2 diluted in human plasma. Measurements were
performed at a bias potential of -150 mV. ..................................................................................... 80
Figure 6-6. Schematic of the alignment device with a lateral resolution of 0.5 mm. .................... 82
Figure 6-7. Exploded view of microfluidic cartridge (left). Photograph of the assembled
microfluidic cartridge (right). Three reservoirs are reduced to one reservoir. The dimensions of
the cartridge are 30 mm × 25 mm. ................................................................................................ 82
Figure 6-8. (a) Schematic of updated microfluidic cartridge design. (b) Microfluidic cartridge
testing with color dye. Blood is added by capillary tube from inlet which contains detection
antibody. (c) The blood sample is pumped toward the outlet, and the mixture of MB and
Ru(NH3)63+ flows to the sensor region. .......................................................................................... 84
Figure 6-9. Photo of the second generation mobile phone biosensing platform. The size of sensing
circuit is minimized to 2 in. × 1 in................................................................................................. 85
Figure 6-10. The calibration curve generated from our POC system using spiked blood samples.
Each data point represents the mean Âą SD of three separate measurements obtained using new
cartridges. ...................................................................................................................................... 86
Figure 6-11. Detection current measurement for evaluating microfluidic cartridge stability at 1
Îźg/mL of PfHRP2 in human blood for 1 month. Each bar represents the mean Âą SD of three
separate measurements obtained using new cartridges. ................................................................ 87
Figure 6-12. Comparison of PfHRP2 measurements of spiked blood samples using ELISA and
the mobile phone device. ............................................................................................................... 88

xiii

Chapter 1. Introduction
1.1 Point-of-care testing
The ability to diagnose patients accurately and promptly is important for subsequent
treatment and disease management. In recent years, diagnostic testing has shifted from
conventional lab-based assays to point-of-care (POC) tests to reduce the time and costs associated
with such testing (Fig. 1-1) [1]. This is especially true for resource-limited countries, which have
limited infrastructure, such as continuous electricity and diagnostic equipment. In these regions,
patients have difficulty accessing centralized medical facilities due to a lack of resources for
transportation, which ultimately hinders treatment. In industrialized countries, such as the U.S., the
time required to run many diagnostic tests is long (> 6 hr), which results in multiple return visits
and missed opportunities for effective treatments [2]. Therefore, providing fast turnaround results
can help medical care providers take appropriate actions and further improve health outcome. POC
assays are designed to be simple, rapid, and low cost, making them well suited for medical
diagnosis. In addition, POC testing can help medical personnel provide proper clinical management,
making it a useful technology for both resource-limited settings and the developed world.

Figure 1-1. Procedures for conventional lab-based diagnostics compared to point-of-care (POC)
testing. Image obtained from [1].

1

A POC device typically employs a biosensor, which is an analytical device for qualitative
or quantitative measurements of biomolecular components using a physical-chemical detector [3].
A biosensor typically consists of three components: bioreceptors, transducer and electronics [4]
(Fig. 1-2). Bioreceptors are biological elements (proteins, nucleic acids) used for sample
recognition, the transducer converts the signal resulting from interaction with analyte to an
electrical signal, and the electronics output the signal to the reader as well as performs any needed
signal processing (amplification, filtering, etc.). Combining the transducer and electronics can
simplify and minimize the size of biosensor. Biosensors can be classified by their physicalchemical properties of the selected transducer, such as thermal, optical, electrical and
electrochemical.

Figure 1-2. Major elements of a typical biosensor. Image obtained from [4].

Lateral flow assays (LFA), such as home pregnancy tests, are one of the simplest and most
common types of POC diagnostic devices. LFAs were first introduced in 1980s [5] for pregnancy
testing and since then, have been applied to other applications, including infectious disease
detection, water/food safety, agricultural contaminants surveillance and drug screening [6]. LFA is

2

a qualitative test for the detection of analytes in biofluids, where the sample is transported via
capillary force on a paper strip and the results are displayed within 5–30 min [7]. LFAs typically
contain a control line to confirm the test results and one or more test lines for the target analyte(s).
While LFAs are valuable for some POC applications, they only provide qualitative results, which
is not useful for diagnostic applications that require biomarker quantification, resulting in the need
for advanced POC technologies.
1.2 Electrochemical biosensors
There are two main detection methods for measuring analytes in biofluids: optical
(fluorescence, colorimetric) and electrical (electrochemical, impedance). Colorimetric is a popular
technique since it is rapid, simple and easy to fabricate. It is a method to determine the concentration
of analyte based on the color change. However, the false positive or false negative signal could be
generated from the background color, which results in an inaccurate measurement. Recently, the
role of electrochemical immunosensors in clinical analysis has increased significantly due to their
high selectivity and sensitivity [8]. More than half of the biosensors applied for detection of
pathogens are based on electrochemistry technology [9]. Since it is low cost, highly sensitive, and
portable, electrochemical biosensors are the most favorable among all sensing techniques.
The sensing principle for an electrochemical biosensor is based on an electrochemical
reaction between a reporter molecule and a substrate solution, which results in the generation of an
electrical signal based on current (amperometric or cyclic voltammetry), potential (potentiometry)
or charge (coulometry) as shown in Fig. 1-3 [3]. Alternatively, electrochemical impedance
spectroscopy (EIS) detects changes in the capacitance and resistance properties of electrode surface
due to analyte binding.

3

Figure 1-3. Process flow of a typical biosensor. Image obtained from [3].

1.2.1 Immobilization of bioreceptors
The immobilization of the bioreceptors is one of the most important aspects for developing
a biosensor. The analytical performance and stability of the biosensor are both heavily influenced
by the immobilization process. There are several methods to immobilize bioreceptors on a sensor
surface, including entrapment, adsorption, covalent bonding, cross-linking and affinity interactions,
as shown in Fig. 1-4 [11].

Figure 1-4. Schematic of main methods for immobilizing bioreceptors on a senor surface. Image
obtained from [11].
4

1.2.2 Principle of electrochemical sensing
An electrochemical biosensor consists of several components, including sensing electrodes,
potentiostat, and a read-out device. A typical electrochemical setup is based on a three-electrode
potentiostat, including a working electrode, reference electrode and counter electrode as shown in
Fig. 1-5 [8,9]. The working electrode is where the reaction occurs, which must be a stable material,
such as carbon or gold. The reference electrode maintains its potential and is used as a reference
for measuring changes in electrical properties of the working electrode. The counter electrode
allows electric charge to flow, thus, there is no current flow between working electrode and
reference electrode, resulting in a stable potential. The voltage between working electrode and
reference electrode is controlled by the potentiostat. An electrochemical process involves a redox
(reduction and oxidation) reaction where electrons are transferred from a chemical substance that
is oxidized to one that is being reduced, and both oxidation and reduction processes occur
simultaneously. For example, reduction is the gaining of electrons from the oxidation process,
which loses electrons. The energy of the electrodes is raised when a negative potential is applied,
and results in electrons required from electrode to solution (reduction current occurs). On the other
hand, the energy can be lowered when a positive potential is applied, which results in an oxidation
current.

Figure 1-5. Example of an electrochemical measurement process for the detection of lactate. Image
obtained from [9].
5

1.2.3 Amperometric biosensors
Amperometry is one of the most popular methods for the development of analytical
biosensors, which is also the most sensitive techniques of all electrochemical sensing methods [14].
In this technique, a fixed potential is applied between the working and reference electrodes, and
the resulting current signal is measured. The magnitude of the applied voltage is based on an
oxidation or reduction generated from electroactive species on the sensor surface. The
concentration of the target analyte in the sample is linearly proportional to the peak value of
generated current. Amperometry usually requires an external potentiostat for generating electrical
signals on the electrodes, as well as an external electronics to read out the results. However,
combining the transducer and electronics together can significantly reduce the overall device size
and improve device performance [15].
Clark et al. demonstrated the first electrochemical sensor for glucose detection based on
oxygen sensors. In this device, glucose and oxygen undergo a redox reaction in the presence of
glucose oxidase [16]. Subsequently, numerous diagnostic applications based on amperometry have
been developed and many tangible applications of amperometry techniques have been developed.
Liu et al. developed an enzyme-free sensor employing nanocubes with graphene as electrocatalysts
for the detection of glucose and H2O2 [17]. The biosensor exhibited a linear response from 0.3 mM
to 3.3 mM, and a lower detection limit of 3.3 ÎźM in buffer solution. Ruecha et al. demonstrated a
paper-based nanocomposite biosensor for the amperometric determination of cholesterol [18]. This
biosensor exhibited a lower detection limit of 1 ÎźM with detection range from 50 ÎźM to 10 mM in
human serum. To achieve a high sensitivity and selectivity, an appropriate bioassay design and
parameter optimization are needed.
There are various methods for fabricating electrodes for electrochemical sensors, including
screen-printing and thin-film deposition. Screen-printing is a popular technique due to its low cost,
simplicity and scalability. Screen-printed biosensors have been developed for oxygen detection
6

[19], protein detection [20,21], blood glucose sensing [22], monitoring metal compositions in
pesticide detecting [23], and water samples [24]. In screen-printing, the desired electrode material
is forced through a custom screen, forming the desired electrode shape on substrate, followed by
air drying. While fast and simple, controlling the uniformity of the electrode surface is challenging
and prone to defects, which can hinder the analytical performance. In thin-film deposition, material
is deposited through a shadow mask onto the substrate via evaporation or sputtering. Although thinfilm deposition is more complicated than screen-printing, it can generate 10× smaller features,
offers better repeatability and improved surface uniformity.
1.3 Microfluidic-based electrochemical biosensors
Most diagnostic tests are performed using bodily fluids (blood, urine, saliva) which
commonly require sample preparation for purifying/concentrating the analytes. In addition, many
assays, such as enzyme-linked immunosorbent assay (ELISA) and reverse transcriptionpolymerase chain reaction (RT-PCR), involve multiple sample processing steps which is time
consuming and laborious. To address these limitations, researchers have utilized microfluidics for
sample preparation and fluidic processing, which offers the advantages of small sample and reagent
volumes (which range from ÎźL to nL), precise fluid control, and fast processing time. For these
reasons, biosensors are integrated with microfluidic components, resulting in lab-on-a-chip
platforms which can offer automated sample analysis in a portable format (typical dimensions of
10-100 mm) [25]. Several microfluidics-based electrochemical sensors have been developed.
Swensen et al. developed a microfluidic biosensor system for continuous measurements of cocaine
in blood [26]. The aptamer-based sensor chip is integrated with target-specific DNA aptamer and
generate electrochemical signal within the microfluidic device.
Silicon and glass were the first materials used for fabricating microfluidic devices. Silicon
was a common material for early development of microfluidic devices due to its compatibility with
microfabrication technologies. The advantages of silicon-based devices include stable surface
7

properties, robust chemical/physical properties and well-known fabrication process. However, it
usually requires expensive facilities for fabrication, therefore, alternative materials have received
more attention [27]. Glass has gained considerable attention for microfluidic devices due to its
chemical stability and optical transparency. It is not permeable and exhibits relatively low nonspecific protein adsorption [28]. However, the fabrication of glass microfluidic devices requires
high temperatures and/or large voltages for bonding glass layers together.
Polymers are another attractive class of materials for microfluidic devices because of their
low cost, ease-of-fabrication, and disposability. Polydimethylsiloxane (PDMS) is a silicone
material, which has been used for fabricating microfluidic devise for many years due to its excellent
optical transparency, biocompatibility, and high compatibility with micromachining processes. The
fabrication of PDMS microdevices is achieved via soft lithography which is a well-known
fabrication process [29]. While PDMS offers many advantages for microfluidic devices, it is not
well suited for some biological applications since its hydrophobic surface can absorb proteins and
its fabrication is not amenable to high volume production.
Recently, researchers have been using plastics [25–27], including polystyrene (PS),
polymethyl methacrylate (PMMA), polycarbonate (PC) polyethylene terephthalate (PET),
polyvinylchloride (PVC) and cyclic olefin copolymer (COC), for microfluidic devices. Similar
with PDMS, plastics are low cost, offer good biocompatibility and amenable to surface
modification while being more mechanically robust, making them better suited for POC diagnostic
devices. While plastics offer several advantages for microfluidic biosensors, more work is needed
to develop simple fabrication techniques and robust surface modification strategies.

8

1.4 Electricity-free microfluidic pumping
1.4.1 Capillary flow
Microfluidic devices usually require external pumps to transport liquids inside
microchannels. These pumping components are bulky, expensive and require external power.
Alternatively, accurate displacement of microfluidic flows can be achieved by using on-chip
micropumps [33]. While micropumps can provide precise, configurable flow rates, they increase the
complexity and costs associated with microfluidic devices. The realization of microfluidic platforms
that are simple, portable and inexpensive requires alternative means for fluid transport without
externally-actuated components. Towards this end, passive pumping is a promising approach that
doesn’t require any moving components and can greatly simplify device fabrication and operation.
There are several methods for achieving passive pumping, such as thermally-driven, and
capillary-driven flows. Among passive pumping methods, one of the most common is capillary
pumping which employs surface tension forces to ‘pull’ liquids through capillaries and
microchannels. This approach also offers simplicity in controlling the flow rate which mainly
depends on the liquid viscosity, geometry, and surface properties of the channels [34]. In particular,
the most important parameter for generating capillary flow is the surface hydrophilicity. Most
polymers and plastics are hydrophobic, which hinders their use for capillary driven flows. While
much work has focused on hydrophilic surface treatments on polymer surfaces, such as chemical
grafting [35], oxygen plasma [36] and UV/Ozone (UVO) treatment [37], further work is needed to
characterize hydrophilic surface treatments on plastics. However, capillary flow suffers from several
disadvantages, such as difficulty in flow control and fluidic processing (e.g. mixing) [38].
1.4.2 Finger-powered pumping
For active pumping methods, electrical power sources are required to generate pressure
head for fluid motion. Recently, an alternative approach for pumping liquids inside microchannels
has been demonstrated based on finger actuation, which is reliable and portable, and eliminates the
9

need for electricity. Various mechanisms for finger-powered pumping have been demonstrated,
such as screw based pumping [39], push-and-release action [40], and vacuum activation [41], as
shown in Fig. 1-6. Based on this principle, the transport of liquids can be easily achieved via finger
actuation. However, valves or diodes are typically required to accurately control the flow direction.
Passive components, such as capillary valves [42], hydrophobic valves [43], or elastomer fluidic
diodes [44], are favorable since they are simple and do not require external power sources. While
functional, these passive components can suffer from poor reliability/reproducibility or require
complicated microfabrication. Therefore, simple finger-powered pumping methods that do not
require valves are desired.

Figure 1-6. Schematic of three types of finger-powered pumping mechanisms: (a) screw-based
pumping, image obtained from [39], (b) vacuum activation pumping, image obtained from [40],
and (c) push and release pumping, image obtained from [41].

1.5 Mobile phone-based biosensors
Mobile phones have been increasingly utilized for biosensing due to their portability,
accessibility and multi-functionality [15]. In particular, mobile phones offer a friendly user

10

interface, fast data processing and wireless communication. A useful component of modern mobile
phones is a high-resolution camera, which has been employed for microscopic imaging. However,
the camera quality varies on different models and brands of phones which can hinder the
consistency of the results. Mobile phones can also be used as a portable power source [45] which
can be used to manipulate liquids in microfluidic chips via electrokinetics. By connecting the phone
with an electrical circuit (i.e. potentiostat), it is also possible to provide electrical signals required
for performing diagnostic measurements. There are two possible electrical connection points on a
mobile phone (audio jack and USB port) that can be used for data transmission. For wireless data
transmission, mobile phones offer several routes, including Wi-Fi, Bluetooth or cellular networks
(3G/4G), which enables rapid transmission of the results in real-time [46]. These capabilities can
also enable for the data to be stored in the Cloud for storing patient medical records.
A smartphone-based fluorescence assay for quantifying albumin in urine was demonstrated
by Coskun et al. [47]. As shown in Fig. 1-7a, this system employs a fluorescence detector affixed
to the phone’s camera, and a transparent tube which contains the urine sample. The fluorescent
signal is processed through an App on the phone and generates the albumin concentration based on
a calibration curve. A lower detection limit of 10 Âľg/mL is demonstrated with a linear range from
0 to 200 Âľg/mL. Although this device is fast and fairly simple, the optical components are bulky
and complicated which limits its usefulness for POC testing. Alternatively, Stemple et al.
demonstrated a POC device for protein quantification based on a microbead immunoagglutination
assay combined with an optic-scatter system used to detect malaria biomarker [48]. Briefly, the
presence of the target antigen within the channel causes the light to scatter, which is captured by
the phone camera as shown in Fig. 1-7b. The concentration of analyte in the sample is determined
based on the intensity of the scattered light. A detection limit of 1 pg/mL was achieved by using
diluting blood, with a detection range from 1 pg/mL to 10 ng/mL. While these results are
comparable to the sensitivity of previously demonstrated POC protein assays, further work is

11

needed to make this system compatible with raw biofluids, such as whole blood. The accuracy and
reliability of optical sensing is greatly influenced by various factors, such as environmental lighting
or sample composition. Therefore, to address this issue, researchers have developed mobile phone
assays based on electric sensors. Lillehoj et al. developed a smartphone POC platform based on
electrochemical detection for rapid protein quantification in human plasma shown in Fig. 1-7c [49].
In the presence of the target antigen, an electrochemical current is generated which is proportional
to the analyte concentration in the sample. The lowest detection limit could be down to 16 ng/mL
with a linear range from 0 to 1,024 ng/mL. Due to its portability, simplicity and capability to
perform analytical measurements in raw clinical samples, this device offers great potential for POC
testing. While promising, this assay requires several sample loading steps and uses an enzymatic
substrate which offers limited room temperature stability. Towards a room temperature-stable assay,
Laksanasopin et al. developed a mobile phone system to perform ELISA for quantifying HIV
protein, as shown in Fig. 1-7d [41]. Dried reagents are stored within a microfluidic cartridge and
become rehydrated by the diluted blood sample, which is transported by negative pressure
generated from a mechanical vacuum spring. The presence of the target analyte causes a change in
the optical density. Therefore, assessments of analytes over a dilution range of 1:128 is
demonstrated in human blood samples. This device could remain stable at room temperature for up
to one month and does not require any external equipment. However, this device only provides a
qualitative result, and the optical signals can be affected by blood samples. Thus, there is a need
for a shelf stable mobile phone biosensor which can generate quantitative and reliable
measurements of protein biomarkers using whole blood samples.

12

Figure 1-7. Examples of mobile phone-based systems for POC protein measurements. (a)
Fluorescent measurement for quantifying albumin in urine. Image obtained from [47]. (b) Opticscatter detection system for measuring protein amount. Image obtained from [48]. (c)
Electrochemical measurement for evaluating malaria protein in human plasma. Image obtained
from [49]. (d) Optical measurement on electrode for quantifying HIV protein. Image obtained from
[41].

1.6 Objectives
Recent reports on mobile phone biosensors have shown that this technology is capable of
rapid, high sensitivity analytical measurements on a portable platform for POC testing. However,
existing devices cannot achieve high sensitivity in whole blood samples or exhibit a limited
detection range. For example, most protein assays, including the gold standard ELISA, exhibit a
detection range from 0.01 ng/mL – 0.1 ng/mL, making it challenging to detect protein markers at
concentrations from ng/mL to Îźg/mL. In addition, many of these assays require multiple samples
loading steps or exhibit poor stability at room temperature. Therefore, the goal of this work is to
develop a finger-powered, room temperature-stable microfluidic electrochemical assay compatible
with a mobile phone-based biosensing platform. This assay will employ an nonenzymatic
electrochemical scheme which can measure protein biomarkers in whole blood samples from 1
13

Îźg/mL to 20 Îźg/mL, encompassing the clinical levels found in patients with P. falciparum infection
[50]. The wide linear detection range of this assay will circumvent the need to run multiple
measurements using diluted samples. To simplify testing, dried reagents will be stored on a
disposable cartridge and reconstituted by phosphate buffer saline (PBS) which is pumped inside
the microchannels via a unique finger-powered pumping mechanism. To achieve this goal, we have
pursued the following objectives:
Objective 1: Investigate UVO treatment on thermoplastics
Objective 2: Explore plastic-based microfluidic device fabrication methods
Objective 3: Optimize electrochemical assay for detection of PfHRP2
Objective 4: Develop a finger-powered microfluidic electrochemical assay
The knowledge and technologies generated from this dissertation represent a culmination of the
experiences that have been obtained throughout my graduate career and will provide researchers
new insights into the development of plastic microfluidic devices for POC testing.

14

Chapter 2. Investigation of UVO treatment on thermoplastics
2.1 Motivation
While plastics offer several advantages for POC microdevices, one of their main limitations
is the lack of robust surface modification strategies, including ones that exhibit long-term stability.
UVO treatment is a simple and effective method for making plastic surfaces more hydrophilic.
Prior reports on the stability of UVO-treated plastics are limited to four weeks [37], which is not
sufficient for applications requiring long-term storage. Therefore, we investigated UVO treatment
on plastics to better understand its usefulness for microfluidic POC applications. In addition, we
present for the first time findings on the long-term stability of UVO-treated plastics for up to 16
weeks and show that the storage condition has a significant impact on the surface stability. We
focused our efforts on PMMA, PC, and COC because these are the most attractive plastics for
biomedical applications due to their excellent optical transparency, biocompatibility and
manufacturability [51]. Surface characterization of UVO-treated and untreated plastic samples was
performed using x-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM) and
contact angle measurements. UVO-treated plastics were also used for protein adsorption studies and
capillary-driven flows experiments.
2.2 Methods and materials
2.2.1 UVO treatment and sample storage
PC and PMMA were purchased from McMaster-Carr (Elmhurst, IL) and COC was
purchased from Zeon Chemicals (Louisville, KY). For XPS and contact angle measurements, COC
and PMMA were cut into 1 cm × 2 cm pieces using a CO2 laser cutter (Universal Laser Systems,
Scottsdale, AZ) and PC was cut using a band saw. UVO treatment was performed using a UVO
cleaner (Novascan Technologies, Ames, IA). For stability studies, samples were treated for 20, 40,
60 or 80 min at room temperature and stored in air, a dehumidified chamber (~20% relative
humidity) or vacuum-sealed plastic bags (Weston vacuum sealer, Strongsville, OH).

15

2.2.2 Contact angle measurements
Static contact angle measurements were carried out using a VCA-2000 video contact angle
analysis system (AST Products Inc., Billerica, MA). 0.5 ÂľL droplets of distilled water were
dispensed onto the samples using a pipette. Contact angle data was acquired at ambient conditions
from three different samples at a minimum of three locations per sample. Each data point is plotted
as the mean of eight separate measurements with outliers removed. New samples were used for
each set of measurements.
2.2.3 XPS analysis and AFM measurements
XPS spectra were obtained using a Perkin Elmer Phi 5400 ESCA system (Physical
Electronics, Chanhassen, MN) at pressures between 10-9 and 10-8 torr, pass energy of 29.35 eV, and
a 45° take-off angle. Elemental composition was calculated from the relative intensities of the C1s
and O1s peak areas obtained from the survey spectral after subtraction of a linear background. AFM
scans were performed under tapping mode in air using a Cypher atomic force microscope (Asylum
Research, Goleta, CA). Samples were imaged using a force constant of 0.2 N/m, scan rate of 2.44
Hz and scan size of 4 µm × 4 µm. Images were processed and analyzed for Root Mean Square
(RMS) surface roughness using Igor Pro software (WaveMetrics, Lake Oswego, OR).
2.2.4 Protein adsorption measurements
Horseradish peroxidase (Px)-labeled anti-PfHRP2 antibody was obtained from
Immunology Consultants Laboratory, Inc (Lake Oswego, OR) and diluted to 10 Âľg/mL in
phosphate

buffered

saline

(PBS)

(Fisher

Scientific,

Pittsburgh,

PA).

3’,3’,5’,5’

tetramethylbenzidine substrate (TMB/H2O2) was obtained from Neogen (Lexington, KY), and
deionized (DI) water (18.3 MΊ-cm) was generated using a Thermo Scientific Smart2Pure water
purification system. PMMA, PC and COC were cut into 2 cm × 2 cm pieces and a 7-mm-diameter
circular well was formed by attaching an adhesive-backed plastic stencil on the surface. Plastic
pieces were exposed to UVO for 20, 40, 60 or 80 min followed by removal of the stencil. 80 ÂľL of
16

1 Îźg/mL antibody solution was dispensed onto each piece and incubated for 30 min, rinsed using
PBS and DI water, and dried under a stream of purified comprised N2. This process resulted in the
adsorption of enzyme-labelled antibodies on the surface. Pieces were then rinsed using PBS and
DI water to remove unabsorbed proteins and dried using N2. 80 ÂľL of TMB/H2O2 was dispensed
onto each sample and incubated for 1 min followed by absorbance measurements at 650 nm using
a UV-visible spectrometer (Shimadzu, Kyoto, Japan).
2.3 Surface chemistry and morphology of UVO-treated plastics
We assessed the surface wettability of UVO-treated plastics by performing static contact
angle measurements PMMA, COC, and PC treated for varying durations (Fig. 2-1). COC and PC
exhibit substantial (60.6% Âą 2.5% and 70.1% Âą 3.2%, respectively) reductions in contact angle with
20 min of UVO exposure, whereas PMMA exhibits a moderate (31.0% Âą 2.9%) reduction.
However, with 40 min of UVO treatment, the contact angle of PMMA is reduced by 54.7% Âą 0.7%,
which is similar to COC and PC. With longer treatment times, the contact angles of COC and
PMMA reach steady state at 60 min and 80 min for PC, which is indicative of the surface becoming
saturated with oxygen-containing species. Among all three plastics, PC exhibits the lowest contact
angle of 10.5º ¹ 0.6° after 80 min of UVO treatment, resulting in a superhydrophilic surface [52].
For PC and PMMA, there is an observable correlation between the treatment duration and surface
wettability where longer times result in lower contact angles. In contrast, UVO treatment times >
20 min had a negligible impact on reducing the contact angle of COC, which is consistent with
findings reported by Bhattacharyya et al. [53]. These results indicate that 20 min UVO treatment
is sufficient in generating a hydrophilic surface on all three plastics, where a more substantial
enhancement in wettability can be obtained for PMMA and PC with longer treatment durations.

17

100
PMMA
COC

Contact Angle (°)

80

PC
60
40
20
0
0

20

40

60

80

UVO Treatment Duration (min)
Figure 2-1. Contact angles of PMMA, COC and PC with varying durations of UVO treatment.
Each bar represents the mean Âą standard deviation (SD) of eight separate measurements.

UVO treatment of plastic results in the generation of an oxide layer on the surface rendering
it more hydrophilic. Organic contaminants on the surface are removed via non-destructive atomic
layer etching by 185 nm and 254 nm ultraviolet light. In the presence of oxygen, 185 nm light
generates ozone while 254 nm light simultaneously excites organic molecules on the surface. The
254 nm light also generates free radicals on the surface that react with oxygen, generating more
oxygen-containing species [54]. The generation of these oxygen-containing species increases the
surface free energy and lowers the hydrophobicity. To study the change in the surface chemistry as
a result of UVO treatment, XPS spectra of plastics exposed to UVO at varying durations were
measured and used to calculate the oxygen/carbon (O/C) ratio on the surface. As shown in Table
2-1, all three UVO-treated plastics exhibit significantly higher O/C ratios compared with untreated
plastics, which is due to the presence of additional oxygen-containing functional groups and
indicative of the formation of an oxide layer. In addition, there is a positive correlation between the
O/C ratio and treatment duration where longer UVO exposure results in a higher O/C ratio since
18

more oxygen-containing species are generated over time. The generation of oxygen species occurs
rapidly within the first 20 min and gradually lessens with longer exposure times as the surface
becomes saturated. After a certain point (~60 min), the oxygen content on the surface reaches a
plateau as indicated by a negligible (< 5%) change in the O/C ratio compared with those of plastics
with longer (80 min) UVO exposure.
Table 2-1. XPS analysis of the O/C ratio on the surface of UVO-treated and untreated plastics
immediately following UVO exposure.
PMMA

COC

PC

0.34 Âą 0.05

0.12 Âą 0.02

0.19 Âą 0.02

20 min

0.48 Âą 0.07

0.42 Âą 0.10

0.47 Âą 0.12

40 min

0.54 Âą 0.09

0.56 Âą 0.06

0.59 Âą 0.04

60 min

0.53 Âą 0.02

0.54 Âą 0.07

0.55 Âą 0.04

80 min

0.56 Âą 0.08

0.55 Âą 0.05

0.56 Âą 0.04

Untreated

UVO-treated

Figure 2-2. AFM tapping mode images of untreated PMMA (a), PC (b), and COC (c), and UVOtreated (80 min) PMMA (d), PC (e) and COC (f). The scan size and z-scale are 4 µm × 4 µm and
50 nm, respectively.

19

AFM scans and surface roughness measurements of UVO-treated and untreated samples
were also obtained to briefly study the influence of UVO treatment on surface morphology as
shown in Fig. 2-2. These results show that UVO-treated plastics exhibit smoother surface profiles
and 2‒7× smaller roughness values compared with untreated plastics as listed in Table 2-2. The
smoother surfaces of UVO-treated plastics are due to the removal of organic contaminants via nondestructive etching from the UVO treatment process.
Table 2-2. Surface roughness values (nm) of UVO-treated and untreated plastic samples
immediately after treatment and after 16 weeks of storage under different storage conditions.

Untreated, Non-stored

PMMA

COC

PC

1.99 Âą 1.34

2.21 Âą 1.64

1.99 Âą 1.57

UVO-treated

Non-stored

0.26 Âą 0.21

0.70 Âą 0.53

0.83 Âą 0.62

(80 min)

Stored in air

2.76 Âą 1.91

2.93 Âą 2.36

2.26 Âą 1.59

Stored in vacuum

0.51 Âą 0.35

1.10 Âą 0.70

1.13 Âą 0.62

2.4 Wetting stability of UVO-treated surface
To better understand the long-term stability of UVO-treated plastics, we performed contact
angle measurements of plastics treated with UVO for 20 min (Fig. 2-3a), 40 min (Fig. 2-3b), 60
min (Fig. 2-3c) and 80 min (Fig. 2-3d) over the course of 16 weeks. For this study, plastics were
stored under different conditions (air, dehumidified and vacuum) to determine its influence on
hydrophobic recovery. Plastic samples stored in air, dehumidified and vacuum conditions are
represented as black, grey and white markers, respectively, in Fig. 2-3. When stored in air, COC
experiences substantial hydrophobic recovery within 4 weeks while PC and PMMA experience
moderate hydrophobic recovery. In general, this trend is consistent for all four treatment times.
When stored in air for > 4 weeks, the degree of hydrophobic recovery increases slightly for all three
plastics, with the exception of COC treated for > 40 min, which steadily increases over time. More
20

importantly, these measurements show that the influence of the storage condition on the
hydrophobic recovery is different for each type of plastic. For PMMA, the storage condition had a
minimal impact on the surface stability for the duration of the study, particularly samples treated >
40 min. This trend is similar for PC, however after 12 weeks, there is a noticeable increase in the
contact angle for samples stored in air. In contrast, the storage condition had a significant impact
on the surface stability of COC where the hydrophobic recovery was dramatically reduced by
storing samples under vacuum. These results indicate that the surface stability can be enhanced by
longer UVO exposure and storing plastics in either a dehumidified or vacuum environment.
XPS spectra of UVO-treated (80 min) plastics stored for 16 weeks under different
conditions were also obtained and used to calculate O/C ratios, which are presented in Table 2-3.
Compared with freshly treated samples (Table 2-1), the O/C ratios for stored samples are lower,
indicating a reduction in the concentration of oxygen-containing species on the surface due to
hydrophobic recovery. UVO-treated plastics stored in air exhibit substantially smaller O/C ratios
by 50% compared with those freshly treated. In contrast, UVO-treated plastics stored in
dehumidified and vacuum conditions exhibit smaller (~10% and ~5%, respectively) reductions in
the O/C ratio. These results are consistent with our long-term contact angle measurements (Fig. 23d) which show that hydrophobic recovery is generally more pronounced when stored in air and
significantly inhibited when stored in dehumidified or vacuum conditions. The slightly higher O/C
ratio exhibited by plastics stored in a dehumidified environment compared with in vacuum is likely
due to the adsorption of water molecules on the UVO-treated surfaces, which reacts with the oxide
layer thus reducing the oxygen content on the surface. Therefore, the oxide layer of UVO-treated
plastics can be maintained by isolating them from moisture-rich environments thereby preserving
the surface hydrophilicity.

21

Figure 2-3. Long-term contact angle measurements of PMMA, COC and PC treated for 20 min (a),
40 min (b), 60 min (c) and 80 min (d) under various storage conditions. Black, grey and white
markers correspond to storage in air, dehumidified and vacuum conditions, respectively. Each data
point represents the mean Âą SD of eight separate measurements.

22

Table 2-3. XPS analysis of the O/C ratio on the surface of UVO-treated and untreated plastics after
16 weeks under different storage conditions.
PMMA

COC

PC

Air

0.27

0.31

0.34

Dehumidified

0.41

0.38

0.38

Vacuum

0.34

0.34

0.36

2.5 Effectiveness of UVO treatment for capillary flow and protein retention
To demonstrate the effectiveness of UVO-treated microchannels for capillary-driven flow,
we performed flow experiments using PMMA microchannels with varying UVO treatment times.
Experiments were also carried out using untreated PMMA microchannels for comparison. 40 ÂľL of
colored liquid was manually dispensed into the inlet of the microchannels using a pipette. As shown
in Fig. 2-4, liquids quickly filled the 80 min-UVO-treated microchannel in < 4 sec while negligible
flow was observed in the untreated microchannel.

Figure 2-4. Capillary flow in untreated (left) and 80 min UVO-treated (right) PMMA
microchannels. Colored dye is dispensed at the inlet of the chips using a pipette.

We also measured the volume flow rate for microchannels with various UVO treatment
times (Fig. 2-5). From this plot, we observe that there is a clear correlation between the treatment

23

time and the flow rate due to the enhanced surface hydrophilicity generated from longer UVO
exposure. This method can be applied to more complex microfluidic networks to develop capillarydriven systems that offer enhanced automation for more simplified operation. These results also
suggest that UVO-treated microchannels can remain effective in pumping liquids via capillary flow
after prolonged periods of storage when properly stored in dehumidified or vacuum conditions.

Figure 2-5. Capillary flow rate as a function of UVO treatment duration for PMMA microchannels.
Inset shows magnified view of the data points at lower (< 40 min) UVO exposure times. Each data
point represents the mean Âą SD of three measurements.

Furthermore, the surface hydrophilicity plays an important role in protein adsorption [55].
To investigate the effect of UVO treatment on protein adsorption, we used a colorimetric detection
scheme based on an enzymatic reaction between an Px-labeled antibody and a chromogenic
substrate. Briefly, a droplet of solution containing Px-labeled antibody was incubated on a plastic
sample for 30 min to allow the protein complex to adsorb onto the surface. The surface was rinsed
using PBS and DI water to remove unabsorbed proteins followed by the application of the substrate
(TMB/H2O2). The substrate reacts with the surface-adsorbed Px to generate a distinct blue color.

24

The absorbance of the droplet was measured at 650 nm using a UV- visible spectrometer, which
correlates with the amount of protein adsorbed on the surface. Absorbance measurements were
performed on PMMA, COC and PC exposed to UVO at varying durations.

0.3
PMMA

Absorbance (a.u.)

0.25

COC
PC

0.2
0.15
0.1

0.05
0
0

20
40
60
UVO Treatment Duration (min)

80

Figure 2-6. Absorbance values of TMB/H2O2 on plastics exposed to UVO at varying durations.
Each data point represents the mean Âą SD of three separate measurements obtained using new
samples.

As shown in Fig. 2-6, longer UVO treatment durations resulted in lower absorbance values
indicating that more hydrophilic surfaces inhibit the absorption of proteins. These findings are
consistent with prior reports on the interactions between proteins and hydrophilic surfaces [56]. Of
the three plastics studied, COC exhibited the lowest absorbance values for all treatment times
indicating the least amount of protein absorption. In contrast, PMMA exhibited the highest
absorbance values, even after 80 min of UVO treatment, indicating more protein adsorption which
is likely due to its surface functional groups [57] and ionic strength [58]. These results indicate that
20 min of UVO treatment is sufficient in lowering the overall amount of protein adsorption on

25

plastic surfaces, where more substantial reductions can be obtained on PC with longer treatment
times.
2.6 Summary
In this chapter, we have presented new findings on the surface chemistry, morphology and
long-term (16 weeks) stability of UVO-treated plastics through XPS, AFM and contact angle
measurements. Specifically, we showed that UVO-treated PMMA, COC and PC experience
hydrophobic recovery within 4 weeks and the rate at which it occurs is dependent on the UVO
treatment duration. Furthermore, we have discovered that the hydrophobic recovery of UVOtreated COC and PC can be inhibited by storing them in dehumidified or vacuum conditions. XPS
measurements revealed that plastics stored in dehumidified and vacuum conditions exhibit higher
oxygen content on the surface compared with plastics stored in air. UVO-treated plastics were also
used for protein adsorption measurements, which showed that UVO treatment minimizes protein
adsorption and this effect is correlated with the treatment duration. Lastly, we demonstrated
capillary-driven flows in UVO-treated PMMA microchannels, which revealed that the flow rate
could be tuned by adjusting the treatment duration.

26

Chapter 3. Plastic-based microfluidic device fabrication methods
3.1 Motivation
Various methods exist for fabricating plastic microdevices with each offering its own
unique set of advantages and disadvantages [59]. Injection molding, one of the earliest and most
common methods, requires expensive machinery and custom dies for each application, making it
poorly suited for low volume production. Recent efforts have focused on alternative fabrication
techniques, including micromilling [60,61], laser micromachining [62,63], stereolithography [64]
and 3D printing [65], which require lower startup costs and can be used for low volume production.
While each of these methods is promising, they are limited to certain types of materials or result in
diminished material properties (e.g. surface finish, optical transparency).
An alternative method that does not suffer from these limitations is hot embossing [66]–
[69]. Hot embossing is a process where a pattern is mechanically stamped into a softened plastic
substrate using a rigid mold. For microfluidics applications, molds are typically fabricated from
metal or silicon using CNC milling or microfabrication techniques, such as photolithography [70],
electroplating [71] and deep reactive ion etching (DRIE) [72]. However, these methods are time
consuming, expensive and require skilled technicians. Other than metal and silicon, alternative
mold materials have been demonstrated on embossed templates, such as PDMS [73,74] and epoxy
[75]–[77]. While these molds can be simpler to fabricate, they tend to suffer from poor
reproducibility for multiple embossing cycles. To address these limitations, we explored the use of
3D printing to generate metal molds for hot embossing plastic microfluidic devices. Through
optimization of the powder composition and processing parameters, we were able to generate
stainless steel molds with superior material properties (density and surface finish) than previously
reported 3D printed metal parts. Molds were used to fabricate PMMA replicas that exhibited good
feature integrity and replication quality. Microchannels fabricated using these replicas exhibited
leak-free operation and comparable flow performance as those fabricated from CNC milled molds.

27

We also investigated the fabrication of microfluidic devices using laser ablation and
adhesive lamination bonding. Various processing parameters (laser power, laser speed, and scan
rate) were characterized for researching laser effectiveness on substrate. Pressure sensitive adhesive
film is utilized for providing a leak-free bonding for microfluidic assembly. With utilization of this
fabrication process, a single microfluidic device can be created within a minute, with similar
performance as devices generated from other fabrication methods.
3.2 Hot embossing using 3D printed molds with thermal bonding
3.2.1 Materials and methods
Microfluidic chips were designed using Siemens NX computer-aided design (CAD)
software and printed using an X1-Lab 3D printer (ExOne, North Huntingdon, PA). This 3D metal
printing system uses a Binder Jetting process (BJP) where a liquid polymer binder is selectively
deposited on metallic powder in a layer-by-layer fashion using a roller. Surface roughness and
replication fidelity measurements of printed molds and embossed replicas were carried out using an
Olympus FluoView FV1000 confocal laser scanning microscope (CLSM). CLSM scans were collected
using a 20× objective (NA = 0.75) scanning at 0.2 μm increments through 120 μm sample thicknesses at
a resolution of 256 × 256 pixels. Surface profile images were generated from stacked CLSM scans and
processed using MATLAB as previously described [78].
Plastic replicas were generated using 1.5 mm-thick PMMA sheets. Prior to molding,
PMMA was rinsed with 2-propanol, deionized (DI) water and dried with compressed N2 gas. Hot
embossing was performed at 1 ton and 120 °C for 30 min using a Carver hot press (Wabash, IN).
The temperature was lowered to 90 °C for demolding. The demolding temperature was carefully
controlled to ensure adequate surface completion and replication quality. Plate alignment was
checked before embossing to ensure uniform force distribution on the PMMA part and the
embossing temperatures were closely monitored throughout the process via a thermocouple.

28

To generate enclosed microchannels, PMMA replicas were thermally bonded to 1.5 mmthick PMMA pieces using a Carver hot press. Through-holes were generated using a CO2 laser
cutter (Universal Laser Systems, Scottsdale, AZ) to create inlets and outlets. Prior to bonding,
PMMA surfaces were treated with UVO for 40 min using a Novascan Technologies UVO cleaner
as we addressed in Chapter 2. In addition to making the surfaces more hydrophilic, UVO treatment
enables bonding to occur at lower temperatures which minimizes the likelihood of channel
collapsing or deformation [54]. PMMA pieces were sandwiched between Al foil to ensure uniform
heat distribution and facilitate alignment in the press. The plates were preheated to 80 °C, followed
by the application of 1 ton for 10 min. The plates were then cooled to room temperature and the
bonded devices were released.
PMMA microchannels were cut normal to their length using a dicing saw (Buehler, Illinois)
and cross-sectional images were captured using a JEOL 6620LV scanning electron microscope
(SEM) at 10 kV with 50× magnification. For flow comparison studies, PMMA microchannels were
also fabricated from stainless steel molds manufactured by CNC milling. The parameters for hot
embossing and thermal bonding were kept consistent for both the 3D printed and CNC milled molds.
Color dyes were dispensed into the inlets of the microchannels using a pipette and fluid flows were
captured using a video camera, which were analyzed for flow rate measurements. Studies to
compare pressure-driven flows in 3D printed and CNC milled molds were carried out by infusing
PMMA microchannels with a solution of 10 Îźm diameter polystyrene beads ordered from Sigma
(St. Louis, MO) in phosphate buffer solution (PBS) at a concentration of 105 particles per mL.
Solutions were infused at a constant rate of 10 μL min−1 using a KD Scientific syringe pump
(Holliston, MA). Flow rates were calculated by tracking the average bead velocity from video
recordings.

29

3.2.2 Fabrication and characterization of 3D printed molds
A major limitation of current metal 3D printing technologies, including BJP, is the
relatively poor surface finish of printed parts. For hot embossing, a smooth mold surface is required
to facilitate the release of the plastic replica without damaging the embossed features. The surface
finish of 3D printed parts can be improved by using very fine powders since thinner deposition
layers can be achieved [79]. However, many 3D printing technologies, including BJP, are constraint
to larger particles due to limitations with the rollers and build bed. To address these issues, we used
a mixture of SS powder with two distinct powder sizes which allows the smaller particles to become
nested in the interstitial spacing of the larger particles, as previously reported [80]. Also, we
discovered that the final part density was significantly improved by incorporating BN powder.
Based on prior experimental studies [80], adding 0.5% wt. BN to the SS powder mixture increased
the final part density to 93% at a lower sintering temperature of 1250°C with negligible distortion.
Photographs and CLSM surface profiles of 3D printed molds with varying concentrations
of BN are shown in Fig. 3-1. The colors in the surface profile images represent the surface height
in the z-dimension. The pure stainless steel mold (Fig. 3-1a) exhibited a surface roughness of 6.61
Âľm Âą 0.71 Âľm. In contrast, molds containing 0.25%, 0.5% and 0.75% BN exhibited improved
surface finishes with surface roughness of 3.83 Âľm Âą 0.53 Âľm, 3.65 Âľm Âą 0.48 Âľm and 3.86 Âľm Âą
0.86 µm respectively (Figs. 3-1b-d). These values are 2× lower than the pure stainless steel mold
and up to 13× lower than previously reported 3D printed metal parts [81], and demonstrate that the
addition of BN greatly improves the surface finish. The mold containing 0.25% BN exhibited
smoother surface and roughness is decreased by 3 times. However, a higher concentration of BN
powder did not seem to have a significance impact on further improving the surface roughness.

30

Figure 3-1. Photographs (upper) and surface profiles (lower) of 3D printed molds containing 0%
(a), 0.25% (b), 0.5% (c) and 0.75% (d) wt. BN. The scan size and z-scale are 600 µm × 600 µm
and 120 Âľm respectively.

3.2.3 Characterization of PMMA replicas
In this work, PMMA with a Tg of 103°C was used. The embossing parameter was fixed at
120°C and 1 ton for 30 min since this setting allows for good surface completion, as previously
reported [82], while minimizing the overall processing time. For assessing mold reproducibility
and physical stability, mold feature integrity, width and height were tracked after multiple cycles.
There is only negligible shrinking (around 2%) in width and less than 1% happened in height after
20 times of embossing cycles.
To assess the functionality of 3D printed molds for hot embossing, we first studied the
surface finish of the replicas. Replicas fabricated from the pure stainless steel mold exhibited a
surface roughness of 8.23 Âľm Âą 0.53 Âľm (Fig. 3-2a) while replicas fabricated from molds
containing 0.25%, 0.50%, and 0.75% BN had substantially lower roughness values of 4.7 Âľm Âą
0.84 Âľm, 4.20 Âľm Âą 0.73 Âľm, and 4.12 Âľm Âą 1.1 Âľm respectively (Figs. 3-2b-d).

31

Figure 3-2. Photographs (upper) and surface profiles (lower) of PMMA replicas fabricated from
3D printed molds containing 0% (a), 0.25% (b), 0.5% (c) and 0.75% (d) wt. BN. The scan size and
z-scale are 600 µm × 600 µm and 120 µm, respectively.

In addition to improving the surface finish, the inclusion of BN also helped to minimize
the presence of defects at the edges of the embossed features that are typically generated during
demolding (Fig. 3-3).

Figure 3-3. Close up images of embossed microfluidic features in PMMA fabricated using the pure
stainless steel mold (a) and mold containing 0.25% wt. BN (b). Hot embossing was performed at
120°C and 1 ton for 30 min. Scale bars, 500 ¾m.

3.2.4 Thermal bonding and the influence of surface roughness
To generate fully enclosed microchannels using embossed components, replicas were
bonded to flat pieces of PMMA. There are various methods for bonding plastics including solvent
bonding [60,61], thermal bonding [84], and adhesive layer bonding [85]. Of these, thermal bonding
32

is a simple approach that can generate a strong and permanent bond. However, thermal bonding
requires high temperatures and forces that can deform or damage plastic features. To mitigate this
issue, a UVO surface treatment was applied to the PMMA surfaces prior to bonding. UVO breaks
down the polymer chains and produces more oxygen-containing functional groups on plastic
surfaces, which improves the overall bonding strength. Therefore, lower bonding temperatures can
be used which helps to preserve the embossed features.
The surface roughness of the embossed part plays an important role in the bonding process.
Therefore, we evaluated the bond quality of PMMA microchannels generated using replicas
fabricated from 3D printed molds. The integrity of the embossed features after bonding was
examined by observing the channel cross section using SEM (Fig. 3-4). Replicas generated using
the pure stainless steel mold at 80°C resulted in incomplete bonding, as evident by the interfacial
gap as shown in Fig. 3-4a. Fully enclosed microchannels could be generated at higher bonding
temperatures (90°C); however, there is evidence of substantial microchannel deformation and
shrinking (Figs. 3-4b). However, leak free bonding could be achieved at 80°C using the mold
containing 0.25% wt. BN with minimal channel deformation (Fig. 3-4c). The difference in the
surface roughness of the embossed parts is quite apparent in regards to bonding microfluidic
devices where replicas fabricated from the mold containing 0.25% wt. BN could be bonded at lower
temperatures with improved microchannel integrity compared with those fabricated from the pure
stainless steel mold.

Figure 3-4. SEM images of cross sections of PMMA microchannels fabricated using the pure
stainless steel mold and thermally bonded at 80°C (a), 90°C (b) and the mold containing 0.25% wt.
BN at 80°C (c). The arrow indicates incomplete bonding. Scale bars, 500 ¾m.
33

3.2.5 Microfluidic device testing
To evaluate the functionality of microfluidic devices embossed using 3D printed metal
molds, PMMA microchannels were fabricated and tested using colored dyes. Due to the UVO
treatment prior to thermal bonding, the inner surfaces of the microchannels were made hydrophilic,
and liquids could be driven inside the channels via capillary flow. We tested two different shaped
microchannels: serpentine and zig-zag. As shown in Fig. 3-5a, colored dyes quickly filled both
microchannels with no observable leaking. To further validate the capabilities of microdevices
fabricated from 3D printed molds, straight microchannels were tested with pressure-driven flows
and visually monitored over time. As shown in Fig. 3-5b, there are no signs of leaking after 1, 5,
10 and 20 min of continuous fluid flow and provides further evidence that these devices exhibit
leak-free operation.

Figure 3-5. (a) Capillary flow in PMMA microchannels fabricated from 3D printed molds. Colored
dye is dispensed at the inlet of the chips using a pipette. (b) Pressure-driven flow inside a PMMA
microchannel fabricated from a 3D printed mold.

Experiments were also performed to compare the flow performance in straight
microchannels fabricated from molds generated using 3D printing and CNC milling, as a
benchmark. We first measured the capillary flow rate in three different channel widths (200 Âľm,
400 Âľm, and 800 Âľm) using DI water. As shown in Fig. 3-6, capillary flow rates for microchannels
34

fabricated from 3D printed molds were comparable to those fabricated from the CNC milled mold.
We also used PMMA microchannels for pressure-driven flows and measured the average flow
velocities. These results are consistent with the capillary flow rate measurements and shows that
there is no significant difference between flow rates in microchannels fabricated from 3D printed
molds and CNC milled molds. These results indicate that plastic microfluidic devices fabricated
from 3D printed molds offer nearly identical flow performance as those fabricated from traditional
manufacturing methods (i.e. CNC milling).

7

45

CNC milled mold

40

3D printed mold

6

35

5

30

4

25
20

3

15

2

10

Flow velocity (mm/sec)

Capillary flow rate (ÂľL/sec)

50

1

5
0
0

200

400
600
Channel width (Âľm)

800

0
1000

Figure 3-6. Comparison of flow performance in PMMA microchannels fabricated from 3D printed
molds (squares) and CNC milled molds (circles). Capillary flow rates and pressure-driven flow
rates are plotted as hollow and solid markers, respectively. Each data point represents the mean Âą
SD of three measurements for capillary flow experiments and five measurements for pressuredriven flow experiments.

While these preliminary results are promising, there are still several limitations with this
approach. First, the lateral resolution is limited to ~200 Îźm. Second, the entire fabrication process
is relatively time-consuming, taking 1 hr to make single plastic device. Therefore, we focused our

35

efforts on using laser cutting to fabricate microfluidic cartridges. The lateral resolution is limited
to 100 Îźm, which is sufficient for this application while offering rapid fabrication.
3.3 Laser ablation with adhesive lamination bonding
While using 3D printed molds to fabricate microfluidic devices via hot embossing offers
several unique advantages, it suffers from a few limitations. First, the lateral resolution is currently
limited to ~200 Îźm. Second, the entire fabrication process is relatively time-consuming, taking 1
hr to make single plastic device. Therefore, we explored an alternative fabrication technique based
on laser ablation and adhesive lamination bonding to fabricate microfluidic cartridges. Recently,
laser cutting/etching has become a popular method for fabricating polymer microfluidic device due
to its speed, simplicity and reproducibility. Machining microfluidic devices with a CO2 laser cutter
on PMMA has been reported for nucleic acid sensing [86]. While laser ablation has limitations with
respect to microchannel lateral resolution (~100 Âľm) and relatively large (Âľm) surface roughness,
it is still a promising technique for prototyping plastic microdevices. Laser ablation of plastics is
typically carried out using a CO2 laser system. Most polymers and plastics leave contamination
residue on the surface and generate toxic gas after cutting. We have investigated the effectiveness
of laser ablation on three plastics including PMMA, PC and COC. PC did not process well with the
laser system as the material showed characteristic charring and yellowing at the processed areas.
COC did not process favorably via laser ablation and resulted in heavy melt accumulation on the
COC surface. Among all types of plastics, PMMA is the most suitable for laser ablation exhibiting
fairly sharp cuts and minimal residue.
3.3.1 Materials and methods
Microfluidic devices were fabricated by etching microfluidic features in PMMA using a
CO2 laser cutter, followed by bonding using double-sided adhesive film (Adhesive Research, Glen
Rock, PA). A schematic of our microfluidic cartridge fabrication process is shown in the Fig. 3-7.
Adhesive film was applied on both sides of middle PMMA layer. Microfluidic features were
36

generated in the middle and top layers using CO2 laser ablation, and then bonded together with an
additional bottom layer.

Figure 3-7. Microfluidic cartridge fabrication process based on CO2 laser ablation and adhesive
lamination bonding.

3.3.2 Optimization of operating settings for laser ablation of PMMA
The shape and size of the ablated features are largely influenced by the operating settings,
such as beam focus, beam profile, exposure time and the power intensity. In order to better control
the ablation process, we optimized various settings. For these studies, we fixed the exposure time
and performed z-axis alignment to ensure that the beam was precisely focused on top of substrate.
We first evaluated the size of ablated feature based on different CO2 power settings (Table 3-1).
The length of the feature was similar under different power setting, and the width slightly increased
with higher power. However, the depth of the feature was significantly increased from 30W to 48W.
At 48W power, the ablated feature was ~4× deeper than that at 30 W. These results indicate that it
is possible to generate high aspect ratio features, with a minimal increase in the lateral dimensions)
simply by increasing the power.

37

Table 3-1. Ablation feature size with respect to different power settings on PMMA. Each
measurement is the average of five different samples.
Power

30W

36W

42W

48W

Length (Îźm)

710

734

737

737

Width (Îźm)

244

248

260

263

Depth (Îźm)

130

296

394

512

We also evaluated the feature size by fixing the exposure time (2 sec) and power (42 W)
and modifying the number of passes of the laser. The results are shown in Table 3-2. For this study,
the PMMA sample was ablated by performing multiple passes of the laser at the same position.
With an increasing number of passes, the feature depth becomes deeper, with doubling with two
passes and tripling with three passes. Interestingly, the width of the feature remained constant with
a slight increase (~ 5%) in the feature length.
Table 3-2. Ablation feature size with respect to number of passes 42W on PMMA. Each
measurement is the average of five different samples.
Power

42W

42W

42W

Number of Passes

1

2

3

Length (Îźm)

737

761

770

Width (Îźm)

260

266

265

Depth (Îźm)

394

694

966

The generation of fully enclosed microfluidic devices typically requires the bonding of
multiple pieces. There are various of factors that need to be considered when bonding plastic pieces
together, such as bonding strength, materials compatible and manufacturability [87]. Several
methods are capable of generating leak-free bonding for polymer microfluidic device, including,

38

thermal bonding, solvent bonding, localized welding and lamination bonding. Lamination bonding
(adhesive bonding) is an attractive technique since it is easy to perform, inexpensive and does not
alter the device features or materials properties. For these reasons, we employed lamination
bonding to fabricate our microfluidic cartridges.
3.4 Summary
In this chapter, we have demonstrated two different approaches for fabricating plastic
microfluidic devices. In the first approach, replicas were generated via hot embossing using 3D
printed molds and bonded together using thermal bonding. Using optimized powder composition
and processing parameters, we were able to generate 3D printed molds with superior material
properties and replication accuracy. Specifically, we show that a SS powder mixture incorporating
0.25% wt. BN powder dramatically improves the part density and surface finish. These
enhancements enable for smoother surface finish of the plastic replicas and improved integrity of
the embossed features. PMMA microchannels fabricated using this method exhibited leak-free
operation with comparable flow performance as microchannels fabricated from CNC milled molds
for both capillary and pressure-driven flows. In the second approach, microfluidic features were
created on PMMA using laser ablation and bonded together using adhesive lamination bonding.
Various operating parameters, including laser power and number of passes, were optimized
enabling for more precise control of feature size, ultimately improving the utility of this technique
for fabricating plastic microdevices.

39

Chapter 4. Electrochemical assay development for PfHRP2 quantification
4.1 Motivation
Immunoassay have widely been used in clinical testing for more than thirty years because
they are the most suitable and effective analytical method [88]. There are several commercial assays
with high sensitivity and accuracy, such as ELISA (Fig. 4-1). However, these assays usually suffer
from complicated sample preparation, and poor stability. For POC test development, it is desired
to develop a room-temperature assay for high sensitivity measurements of protein biomarkers in
whole blood samples. To achieve this, we investigated two different detection schemes and studied
various assay parameters, including the immobilization of bioreceptors, concentration of the
antibodies/reagents, optimization of substrate and applied potential.

Figure 4-1. Schematic of sandwich ELISA. Target protein is encapsulated between two layers
antibodies, followed by TMB reacting with Px and generate signal. The signal should be
proportional to the concentration of target protein. Image obtained from [88].

40

4.2 Horseradish peroxidase-based electrochemical assay
A schematic of the PfHRP2 electrochemical assay is shown in Fig. 4-2. Anti-PfHRP2 IgM
is immobilized on the working electrode (Fig. 4-2a). In the absence of PfHRP2 antigen, the Pxconjugated detection antibody is washed away during the rinsing step and a negligible signal is
generated. However, when PfHRP2 is present in the sample, it binds to sensor surface (Fig. 4-2b)
allowing surface immobilization of the Px-conjugated detection antibody (Fig. 4-2c). Upon
application of a bias potential, an electrochemical reaction takes place between Px and substrate
(i.e. TMB), resulting in a high signal (Fig. 4-2d). Based on this scheme, the concentration of
PfHRP2 in the sample is proportional to the current generated during the electrochemical reaction.
In this chapter, we are introducing various bioassay development methods and evaluate their
performance.

Figure 4-2. Schematic of an enzymatic PfHRP2 electrochemical assay. (a) The sensor is
immobilized with capture antibody on surface. (b) If present, sample binds to the capture antibody.
(c) Detection antibody is immobilized on the sensor via binding to sample, forming a sandwich
structure. (d) Peroxidase reacts with substrate generating a current that is proportional to the
PfHRP2 concentration.

4.2.1 Materials and methods
Gold sensor chips were purchased from GeneFluidics (Irwindale, CA). Amperometric
measurements were performed using a multichannel electrochemical workstation obtained from
41

GeneFluidics. Before electrochemical testing, gold sensors were carefully examined for defects,
such as debris and scratches. Anti-PfHRP2 mouse IgM (capture antibody) and Px-conjugated antiPfHRP2 mouse IgG (detection antibody) were obtained from Immunology Consultants Laboratory
(Portland, OR). Recombinant PfHRP2 was purchased from CTK Biotech (San Diego, CA).
Samples were prepared by serially diluting PfHRP2 antigen in PBS, human plasma or whole blood
(Bioreclamation Inc., Westbury, NY). Bovine serum albumin (BSA), phosphate buffer solution
(PBS), PBS-casein, hydrogen peroxide (H2O2), 4-Aminophenol (4AP), dimethyl sulfoxide
(DMSO), Traut’s Reagent (2-iminothiolane) and ethylenediaminetetraacetic acid (EDTA) were
purchased from Sigma. DI water was generated using a Thermo Scientific Smart2Pure water
purification system (Tustin, CA).
For analyte measurements, 40 ÎźL of capture antibody was immobilized onto the sensing
region using a pipette, and incubated on the sensor for 1 hr followed by rinsing with PBS and drying
with N2 gas. The antibody-immobilized sensor was incubated with 40 ÂľL of PBS-casein for 30 min
at 25ÂşC to block the surface from nonspecific binding. The sensor was then rinsed with PBS, dried
using N2 gas, followed by adding 40 ÂľL sample and 40 ÂľL detection antibody and substrate
sequentially. Next, the application of a −200 mV bias potential was applied on the working
electrode after 1 min. All electrochemical measurements were performed at room temperature
under ambient conditions using new sensors.
4.2.2 Antibody immobilization optimization
Protein immobilization is an important step for preparing the biosensor surface since the
amount of immobilized protein affects the sensitivity and detection range. There are several
immobilization methods to attach proteins onto a sensor surface, which can be grouped into three
main categories: physical adsorption, covalent bonding and bioaffinity interactions [62,63] as
shown in Fig. 4-3. Since random orientation and structural deformation leads to a loss of protein
activity, a suitable immobilization is required to retain biological activity. We have assessed
42

different immobilization methods to determine the one that offers the optimal detection
performance.

Figure 4-3. Different methods for performing immobilization and functional groups used for
orienting antibodies on substrate [62]

4.2.2.1 Physical absorption
Proteins can be absorbed on surfaces via intermolecular forces, mainly through
electrostatic attraction, and hydrophobic and polar interactions. As a result, proteins are likely to
be heterogeneous and randomly oriented on the electrode. The main drawbacks of this method are
weak attachment resulting from the random orientation which may result in proteins being washed
away during rinsing.
The preliminary results shown in Fig. 4-4 were generated using sensors employing physical
adsorption for the immobilization of capture antibodies. Briefly, we incubated capture antibodies
on the sensor for 2 hr, sequentially added antigen (PfHRP2) serial diluted in PBS from 16 ng/mL
to 1,024 ng/mL, added detection antibody conjugated with Px, followed by the addition of
TMB/H2O2 substrate for electrochemical reaction (TMB was first selected as our substrate and we
will further discuss it in the following sections). The resulting electrochemical signal was acquired

43

by applying a -200 mV bias potential. As shown in Fig. 4-4, the analyte concentration is
proportional to the current up to 256 ng/mL.
-2,500

Current (nA)

-2,000
-1,500
-1,000
-500
0
0

16

32
64
128
256
512
PfHRP2 concentration (ng/mL)

1,024

Figure 4-4. Amperometric signals of PfHRP2 diluted in PBS using sensors prepared by physical
absorption immobilization technique. Each bar represents the mean Âą SD of four measurements.

4.2.2.2 Amine group covalent binding
For the physical absorption immobilization method, it is difficult to control the antibody
orientation, resulting in randomly-oriented antibodies on the surface. In a sandwich ELISA scheme,
the antigen only binds to a single epitope on the capture antibody. Therefore, if the epitope is not
oriented upward, the probability for capturing the target antigen is lowered. In addition, the forces
acting between the antibody and the substrate mainly depend on low-energy interactions, which
results in weak antibody binding strength to the substrate. Multiple sample-pipetting and residuewashing steps can destabilize antibody attachment, so this approach is not practical for ELISAbased assays.
Therefore, an alternative covalent binding method was studied. Covalent attachment
immobilization is a frequently used method in microfluidic assays [89]. The activated surface reacts
44

with amino acid residues on the protein to form a stable linkage. In this approach, antibodies are
covalently bound to the gold surface, as shown in Fig. 4-5. The chemical bonds are usually
generated by reagent treatment, such as chemical self-assembly and plasma treatment [91].
Carboxyl, amine and thiol group are typical chemical bonds used for antibody attachment.
Carboxyl-based covalent bonding on gold has been used for many years for the immobilization of
proteins [92]. For carboxyl self-assembly on gold, 11-Mercaptoundecanoic acid (11-MUA) is
applied to the surface for 2 hr at 10 mM (these conditions are selected based on optimization). For
activating the carboxyl groups, [1-ethyl-3-(3-dimethylaminopropyl) carbodiimide] (EDC) in
combination with N-hydroxysuccinimide (NHS) at a 2:1 ratio was applied on to gold substrate and
incubated for 1 hr. Capture antibody was then added, which linked with the gold electrode via
amine chemical bonding. The spiked sample and substrate were subsequently added to the surface
to generate an electrochemical reaction. Electrochemical measurements were taken with a potential
bias at -200 mV, and the results are shown in Fig. 4-6.

Figure 4-5. Schematic of 11-MUA self-assembly on gold electrode and subsequent immobilization
of capture antibody and antigen. Image obtained from [93].

45

-2,500
Physical absorption

Amine covalent binding

Current (nA)

-2,000
-1,500
-1,000
-500
0
0

16

32
64
128
256
512
PfHRP2 concentration (ng/mL)

1,024

Figure 4-6. Amperometric signals of PfHRP2 diluted in PBS via amine- group covalent binding
immobilization technique. Each bar represents the mean Âą SD of four measurements.

Covalent binding exhibited stronger bonding compared to physical absorption and
prevented enzyme release into the reaction environment. Absorption binding resulted from weak
non-specific forces, such as van der Waals, hydrophobic attraction, and hydrogen bonds which are
weaker than chemical covalent bond. In addition, covalent binding method increases enzyme
activity when couple with substrate supports [94]. Even though there are two more steps for primary
antibody immobilization, the amine covalent binding method is considered to facilitate long-term
storage and reusability of immunosensors [95]. Based on our preliminary results shown in Fig. 46, the amine binding method showed similar signals from 0 to 1,024 ng/mL as the physical
adsorption method. However, amine groups generated higher currents at higher PfHRP2
concentrations above 512 ng/mL. These results indicate that there is more capture antibody
immobilized on the sensor and has the ability to capture more analyte in the sample, therefore
generating a higher signal.

46

Another parameter that we investigated was the incubation time. For traditional ELISA, it
requires 6 hr to perform a single measurement. It is desirable to develop a faster diagnostic test
which can provide prompt results to patient and medical technicians. In order to achieve this goal,
various techniques have been explored to reduce the incubation times, including the use of
nanomaterials [96], magnetic stirring [97] and electrophoresis [98]. Alternatively, the incubation
time of reagent preparation can be dramatically reduced via electrokinetic-assisted protein transport
[68,69]. Specifically, applying a bias potential on non-uniform electrodes generates electrokinetics
forces which transports biomolecules closer to the surface, resulting in faster protein recognition
and binding. We employed this technique for our assay and optimized the electric field parameters,
resulting in an incubation time of 5 min which is 12× faster than ELISA. PfHRP2 recognition was
achieved by applying 400 mV for 3 min, and detection antibody binding step was assisted by a 200
mV for 1.5 min. The performance of this technique was evaluated by measuring buffer samples
spiked with PfHRP2 from 0 ng/mL to 1,024 ng/mL compared with an incubation-based approach.
-3,500

Without electrical

With electrical

-3,000
Current (nA)

-2,500
-2,000
-1,500
-1,000
-500
0
0

16
32
64
128
256
PfHRP2 concentration (ng/mL)

512

1,024

Figure 4-7. Amperometric signals of PfHRP2 diluted in PBS obtained from electric field-assisted
sensors (solid) and sensors undergoing incubation. Each bar represents the mean Âą SD of two
measurements.

47

As shown in Fig. 4-7, the signals for electrical field-assisted sensors are (1.3×) higher than
the signals obtained using a standard incubation method. Based on these results, electrical fielddriven protein recognition/binding not only minimizes the testing time but also results in improved
signal-to-background ratio (SBR) and less incubation time.
Measurements of PfHRP2 at higher concentrations were performed in PBS, plasma and
whole blood to evaluate the detection range of this assay. As shown in Fig. 4-8, measurements from
only PBS matrix exhibited a linear relationship. In general, signals in whole blood are slightly
higher than in plasma samples due to matrix effect. However, human plasma samples exhibited
better SBR than whole blood samples, therefore detection limit of whole blood is constrained in
this bioassay. Human blood contains lipids which can bind to the antibody-coated sensor and
interfere with protein recognition and binding [101]. In addition, red blood cells contain
hemoglobin which exhibits pseudo peroxidase activity [102], which can interfere with the
electrochemical reaction if the immune complex is not well washed. Furthermore, the detection
signals in both blood and plasma were substantial lower compared with PBS and further work will
focus on improving the assay performing in biofluids.
-1,200
PBS

Plasma

Blood

Current (nA)

-1,000
-800
-600
-400
-200
0
0

16

64
256 1,000 4,000 8,000 16,000
PfHRP2 concentration (ng/mL)

Figure 4-8. Amperometric signals of PfHRP2 diluted in PBS (black), human plasma (striped) and
human blood (gray). Each bar represents the mean Âą SD of four measurements.
48

4.2.2.3 Thiol group covalent binding immobilization method

Gold
Figure 4-9. Schematic of the working mechanism for thiol group immobilization method. Image
modified from [103].

We investigated another approach for antibody immobilization based on thiol-group
interaction to simplify the immobilization protocol. Using this approach, the 11-MUA incubation
and EDC/NHS activation steps can be eliminated from the protocol. The capture antibody is
processed by Traut’s Reagent forming thiol chemical bonds, as shown in Fig. 4-9. Thiolated
antibody solution was prepared by combining 990 ÂľL of PBS (Ethylenediaminetetraacetic acid
(EDTA)) containing of 100 Âľg/mL anti-PfHRP2 IgM with 10 ÂľL of dimethyl sulfoxide (DMSO)
(1 mg/mL) in Traut’s Reagent, incubating at 25ºC, and centrifugation to remove residual DMSO.
Precipitated antibody was reconstitute using PBS and incubated on the gold sensor for 1 hr. This
immobilization process can be completed in 2 hr with thiol chemical bonds attached to surface. The
antibody-immobilized sensor was incubated with PBS-casein mixed with 30% of stabilizer to block
the surface from nonspecific binding.
The minimized processing steps for primary antibody immobilization improved the
simplicity and convenience of the sensor preparation protocol. Thus, we spiked PfHRP2 in human
blood from 0 ng/mL to 16,000 ng/mL and performed the measurement for comparing the result
with amine group immobilization method. As shown in Fig. 4-10, the results indicated that the thiol

49

immobilization method exhibited higher signals and offered a better SBR difference. Therefore, we
selected thiol immobilization method as our bioassay development method.
-1,000

Amine group

Thiol group

Current (nA)

-800
-600
-400
-200
0
0

16

64
256 1,000 4,000 8,000 16,000
PfHRP2 concentration (ng/mL)

Figure 4-10. Amperometric sensing of PfHRP2 diluted in human blood based on carboxyl group
antibody immobilization (solid) and thiol-group antibody immobilization (striped). Each bar
represents the mean Âą SD of two measurements.

4.2.3 Substrate optimization
In our enzyme-based electrochemical scheme, Px was selected as our reporter conjugated
to the detection antibody. TMB is the most common chromogen used for Px and evaluated as the
electrochemical substrate [104]. We initially selected TMB as the substrate for Px in ELISA, which
usually develops a soluble blue reaction product that may be read at 370 or 655 nm via spectrometer.
Moreover, measurement of Px activity can be achieved by using amperometric method [25,71].
The Px-TMB working mechanism is shown in Fig. 4-11, TMB reacts with Px and generates a redox
cycling with gold electrode [97]. The value of the signal is proportional to the amount of Px on
detection antibody. The preliminary results presented above were all obtained using TMB.

50

Figure 4-11. Schematic of the detection scheme for PfHRP2 detection based on a redox reaction
between an enzyme reporter and TMB+H2O2. Image obtained from [97].

For ELISA and many colorimetric immunoassays, TMB is used as the substrate. However,
TMB is a light and temperature sensitive chemical, and easily degrades without proper storage. The
need for refrigeration and light-protected storage hampers its practicality for POC applications.
Also, the detection signal at higher PfHRP2 concentrations is low, therefore, we investigated ways
to increase the amperometric signals. There are several ways for enhancing the detection signal in
electrochemical assays, such as utilizing nanomaterials for increasing the reaction area and speed,
optimizing the immobilization method for capturing more target on the surface, or selecting a
mediator for enhancing redox reaction. All these methods are useful, however, considering the
simplicity and stability approach, we focused our efforts on substrate optimization.
4AP is an organic compound which can be used for electrochemical reactions and exhibits
higher sensitivity due to its distinct redox reaction. In each electrochemical reaction, there are two
electrons transfer involved for 4AP redox reaction, which is 2× as much as the TMB reaction [105].
Thus, 4AP will promote electron transportation speed, and generate more current than TMB.
Furthermore, TMB as a substrate for Px is more sensitive than other chromogens, which can be
degraded by sunlight or fluorescent lights, resulting in poor signal intensity [106], and may produce
51

false positive signal affecting the accuracy of measurements. Additionally, 4AP can be stored as a
solid form making it possible for storage at room temperature without decay; it’s a great advantage
for applying 4AP as our device substrate. Future work will focus on characterizing the room
temperature stability of 4AP.
As shown in Fig. 4-12, there is a linear relation up to 20 Âľg/mL using 4AP compared with
TMB which remains flat from 1 Âľg/mL to 20 Âľg/mL. Based on these preliminary results, the use
of 4AP as substrate provides us a wider detection range and higher SBRs. These results are
consistent with previous reports on achieving lower background current level in Px via catechol/obenzoquinone redox couple [107].
25
TMB

4AP

20

SBR

15
10
5
0
0

1
10
PfHRP2 concentration (Âľg/mL)

20

Figure 4-12. Amperometric results were obtained from spiking sample in PBS. Signal-tobackground ratio was used for comparing different substrate of 4AP and TMB. There is negligible
signal difference for TMB for concentration above 1 Âľg/mL. Each bar represents the mean Âą SD
of two measurements.

For amperometric measurements, the value of current is evaluated based on a constant
applied potential. Selection of optimal potential for the enzymatic reaction (Px-TMB) is significant.
Cyclic voltammetry (CV) was conducted from +200 mv to -600 mV to find the optimal signal for

52

SBR. The potential range for the reduction reaction occurred between -100 mV to -300 mV. We
tested each potential with 50 mV increments, and -200 mV and -150 mV generated the highest
SBR. Therefore, PfHRP2 sample was spiked in human plasma samples at varying concentrations
from 0.1 Âľ/mL to 50 Âľg/mL, and the two bias potentials were applied to evaluate their signal
generation. As shown in Fig. 4-13, -150 mV resulted in improved linearity up to 20 Âľg/mL.
Therefore, we selected -150 mV as the bias potential for electrochemical measurements.
30
-150 mV

- 200 mV

25

SBR

20
15
10
5
0

0

100

500
1,000 5,000 10,000 20,000 50,000
PfHRP2 concentration (ng/mL)

Figure 4-13. Amperometric signals of PfHRP2 diluted in human plasma using a bias potential of 200 mV (solid) and -150 mV (striped). The scheme is based on thiol immobilization method using
4AP as the substrate. Each bar represents the mean Âą SD of two measurements.

4.3 Methylene blue-based electrochemical assay
We have developed new approaches for detecting PfHRP2 by using Ru(NH3)63+, which
relies on methylene blue (MB) as electron mediator for increasing electron transfer within
electrochemical redox cycling as shown in Fig. 4-14. The utilization of a mediator assists the
electron transfer within electrochemical redox cycling [108]. MB is a well-known electrochemical
properties, which has been used as an electron transfer mediator within in redox cycling [105,106].
The electron transfer scheme is formed by Ru(NH3)63+/Ru(NH3)62+ redox cycling at the outer sphere
53

and MBox/MBred react at the inner surface. When a negative potential is supplied from the power
supply, Ru(NH3)63+ is reduced to Ru(NH3)62+ which reacts with MBox and oxidizes back to
Ru(NH3)63+, the electron transfer process for EC redox cycling is enhanced due to regeneration of
Ru(NH3)63+ for each redox cycling.

Figure 4-14. Schematic of electrochemical redox cycling based on the redox reaction between MB
and Ru(NH3)63+ .

4.3.1 Materials and methods
For this bioassay development, we utilized the same immobilization method as addressed
in Section 4.2.2.3. For the selection of substrate, hexaamineruthenium(III) chloride (Ru(NH3)63+),
MB, phosphate-buffered saline (PBS, pH 7.4) and other reagents for buffer solutions were obtained
from Sigma-Aldrich (St. Louis, MO). StabilBlockÂŽ Immunoassay Stabilizer was obtained from
SurModics, Inc.

54

4.3.2 Antibody immobilization optimization
With the development of this sensor, the protocol we used was identical to what was
described in Section 4.2.2.3. The capture antibodies are thiolated with Traut’s reagent for 1 hr
incubation, followed by immobilized thiolated antibody on gold sensor for 1 hr. Then, we applied
a surface blocker, which is a stabilizer containing with PBS-casein for 30 min. After washing away
the blocker, compressed N2 was used to blow away the residues on the gold sensor, which were
stored in a dehumidified chamber before assembly with the other pieces. For the negative control
sensor, capture antibody was not immobilized on the surface. However, the control sensor was
coated with a blocker to prevent PfHRP2 binding. The negative control sensor will generate a
background signal, which can be used to determine a false measurement. In the immunoassay
electrochemical scheme we proposed, the current value for control sensor is always higher than the
testing sensor. With this control sensor, we can validate the accuracy of the test by comparing the
result with the background signal generated by the control sensor. Therefore, the quality and
reliability of the POC diagnostic can be improved.
4.3.3 Substrate optimization
The bioassay detection scheme is illustrated in Fig. 4-15. The electron transfer rate is faster
when there is no PfHRP2 on the electrode surface and results in a higher signal. When the target is
bound to the surface, the detection antibody binds to the sensor surface, effective forming an
insulating layer. This insulating layer is proportional to the concentration of the protein layer
formed on surface and reduces electron transfer with the electrode. However, the detection antibody
will be washed away if there is no target on the surface. Then, there is no obstacle for electron
transfer when there is no protein insulating layer is formed on the surface. Based on this scheme, a
sample without the target analyte should generate a large current, which becomes reduced with
increasing concentrations of PfHRP2 in the sample.

55

Figure 4-15. Schematic of electron transfer scheme on different antigen–antibody binding
conditions. (a) When PfHRP2 is presented on surface, the electron transfer is inhibited. (b) When
PfHRP2 is absence, the electron transfer is higher.

4.3.4 Electrical field preparation optimization
For the recognition of antigen binding on epitope of antibodies, it usually requires 1 hr for
completing antigen-antibody binding in the corresponding surface. It is reported that the analyte
transport rate can be accelerated by the production of an electrical field near the electrode surface
[111]. With the electric field-driven incubation steps, it reduces the time-consuming incubation
step to 10 min, while also enhancing the assay sensitivity.
In this work, we applied an electric field for assisting the antigen and detection antibody
recognition steps. The transport of charged proteins is highly affected by the electrical field
generated from electrode surface. It is reported that the more accelerated immunocomplex will be
increased by the positive driving potential [100]. We have demonstrated a similar test by various
potential of 200 mV, 400 mV, and 600 mV with the same driving time, followed by adding a
56

substrate for generating the redox cycling. The results showed that 400 mV exhibited the lowest
signal, resulting from the most protein layer formed on the surface. However, 600 mV exhibited
worse performance than 400 mV since a higher driving potential results in more electrostatic
repulsion causing the immunocomplex to be slower than the transport rate.
The potential driving period was as investigated. We have tested the different potential
durations from 1 min to 5 min at 400 mV, followed by Ru(NH3)63+ and MB reaction. As shown in
Fig. 4-16, the lowest signal was generated at 3 min, which indicates that there is the most protein
layer developed on the surface. Interestingly, the 4 min driving time exhibited the similar signal to
background performance as 3 min driving time. However, it is favorable to apply a shorter driving
potential time, thus, 3 min of driving time at 400 mV driving potential was selected for antigendetection antibody complex formation on the sensor region.

-3,000
No Target

Target

Current (nA)

-2,500
-2,000
-1,500
-1,000
-500
0
1 min

2 min
3 min
4 min
Electrical Field Duration

5 min

Figure 4-16. Effect of different duration of 400 mV driving potential on whole blood samples
containing 10 Îźg/mL for comparing the value of target and without target. Each bar represents the
mean Âą SD of three separate measurements obtained using new sensors.

57

To validate this electric field incubation concept for protein recognition, measurements of
human blood spiked with 10 Îźg/mL PfHRP2 were performed. Measurements were also performed
without electrical field preparation for comparison. As shown in Fig. 4-17, the electrical field
incubation group exhibited a lower voltammetric detection current, resulted from a larger amount
of protein layer forming on the surface. Conversely, there is less protein layer on the surface without
an electric field driven, resulting in a higher detection current.

Figure 4-17. Cyclic voltammograms of whole blood samples containing 10 Îźg/mL with electrical
field preparation (EFP) and without EFP using open well chips.

To validate the proposed scheme on the microfluidic cartridge, measurements of human
blood spiked with PfHRP2 were also performed. As shown in Fig. 4-18, the 10 Îźg/mL PfHRP2
sample exhibited a lower voltammertric current signal since the electron transfer rate was inhibited
by the protein formation on the substrate. However, the sample without PfHRP2 showed a higher
current value due to the negligible protein layer presented on surface and resulted in a faster electron
transfer rate.

58

Figure 4-18. Cyclic voltammograms of whole blood samples spiked with 10 Îźg/mL PfHRP2 and
without PfHRP2. The measurement is performed using a microfluidic cartridge.

4.4 Summary
In this chapter, we characterized and optimized the design and performance of an
electrochemical assay for PfHRP2 detection. Our initial efforts were focused on an ELISA-based
detection scheme using Px and TMB as the reporter and substrate, respectively. We investigated
different immobilization methods, including ones based on adsorption, amine covalent binding and
thiol covalent binding. Covalent binding resulted in a wider detection range compared with
adsorption, however, amine covalent binding is complicated and time-consuming. Thiol covalent
binding offered improved performance compared with amine covalent binding while being faster
and simpler. We also evaluated two substrates, TMB/H2O2 and 4AP/H2O2, and the use of electrical
field-assisted protein recognition/binding which improved the assay stability and reduced the
processing time, respectively. We then focused our efforts on a MB-based electrochemical assay

59

to enhance assay stability for long-term storage. With the nonenzymatic reagent and stabilized
antibody, the proposed immunoassay exhibited higher stability than the first generation assay.
Reagents are dried and protected by optimized stabilizing condition, the cartridge stability can
maintain for a month at room temperature with minimal signal loss.

60

Chapter 5. Finger-powered microfluidic pump
5.1 Motivation
Microfluidic pumping based on finger actuation is an attractive method due to its simplicity
and ease of fabrication. Prior finger-powered pumps are based on complicated fabrication
processing or require additional fluidic components (valves, diodes) which can suffer from poor
reproducibility [93–95]. To address these issues, we have explored the development of a valveless,
piston-based micropump, which is robust and can easily be incorporated into plastic microfluidic
devices.
5.2 Materials and methods
5.2.1 Finger-powered pump fabrication
PDMS prepolymer and curing agents (Sylgard 184) were obtained from Dow Corning
(Midland, MI). PDMS was fabricated by mixing PDMS prepolymer and curing agents, degassed
and poured onto the molds, and cured for 2 h at 80 °C. Individual PDMS pieces were cut by using
CO2 laser cutter. 1.5 mm-thick PMMA and rubber gaskets were obtained from McMaster-Carr
(Elmhurst, IL). Prior to fabrication, PMMA was rinsed with 2-propanol, deionized (DI) water and
dried with purified N2 gas. Capillary tubes were purchased from Sigma (St. Louis, MO) and doublesided adhesive film was purchased from Adhesive Research (Glen Rock, PA). Fluorescent dye for
examining washing efficiency within microchannel was ordered from ThermoFisher (Waltham,
MA), and the image was captured by fluorescent microscope with absorption/emission wavelength
at 347/448 nm. Colored dyes for flow visualization were purchased from a local vendor.
5.2.2 Finger-powered pump characterization
Flow experiments were performed using microchannels having a height of 100 Âľm, length
of 15 cm and varying widths (100 Îźm, 200 Îźm, 400 Îźm, 600 Îźm, and 800 Îźm). 80 ÎźL of colored
dye was added to the reservoir prior to bonding the top PMMA layer. The flow was initiated by
placing metal weights on top of the piston and recorded using a video camera. The flow rate was
61

calculated from still frame images of the video recordings by measuring the distance of the flow
front at specific time intervals. We compared our experimental results with theoretical calculations
of 2D pressure-driven flow inside a rectangular microchannel.
Measurements of the actuation force were performed using an M&A Instruments digital
push-pull force gauge. The force gage was mounted on a vertical stage and positioned so that the
compression rod was aligned with the piston of the cartridge (Fig. 5-1). The force gauge was
connected to a PC and measurements were recorded using the manufacturers’ software.
Measurements were performed using microchannels having various widths filled with colored dye.

Figure 5-1. Photograph of experimental setup for piston actuation force measurements. The
compression rod is aligned with the piston, and the resulting data is transmitted to a PC. The inset
shows a close-up view of the compression rod pressing the piston during a measurement.

5.3 PDMS membrane-based pump
The first generation of our finger-powered pump is shown in Fig. 5-2. It consists of two
layers of PMMA with a single PDMS layer in the middle, which were bonded together by adhesive
film. A positive pressure was applied on the PDMS membrane by depressing it with a finger,
resulting in the liquid being pushed into the microchannel. Upon release of the membrane, we

62

observed that some of the liquid flowed back into the reservoir due to the negative pressure
generated from the recovering motion of the membrane. Thus, we investigated methods to fabricate
one-way valves inside the microchannel to prevent backflow (Fig. 5-3). For the fabrication of the
PDMS valve, a 100-Âľm thick PDMS flap was etched using a CO2 laser cutter.

Figure 5-2. Photograph of the PDMS membrane-based pump before (left) and after(right) finger
actuation.

Figure 5-3. (a)(b) Schematic of the working principle of the fluidic valve to regulate flow. (c) Top
view the one-way valve captured by optical microscopy. Scale bar, 500 Îźm.

63

We integrated three reservoirs into a single cartridge and performed flow testing using
colored dye as shown in Fig. 5-4. Although the valve was capable of preventing backflow and
regulating the flow direction, it exhibited some limitations. First, the liquid could not fully cover
the sensor region, thereby hindering the accuracy of electrochemical reaction. Second, liquid
leaking occurred frequently during the third activation step due to weak bonding strength between
PDMS and PMMA. Last, the alignment between the PDMS and PMMA layers is complicated,
especially for the valve’s fabrication which requires precise alignment. Due to these issues and
difficulties, we investigated alternative pump designs.

Figure 5-4. Microfluidic cartridge testing using color dyes. (a) The green dye was dispensed
toward outlet. The arrow indicates poor liquid coverage for sensor region. (b)(c) Red and green
were dispensed sequentially. The arrow reveals blue dye leaking resulting from incomplete bonding.

5.4 Piston-based pump
To address the limitations of the PDMS membrane pump described above, we employed a
rigid piston as the transducer element. The design is similar to a mechanical piston, where a rigid
piston slides inside a hollow cylinder, as shown in Fig. 5-5. Three different piston designs were
investigated. In the first design, the high friction generated between the PDMS and PMMA resulted
in the PDMS on the outer layer break, as indicated by the red circle in Fig. 5-5a. In the second
design, a thinner PDMS membrane was placed on top of a PMMA notch, as shown in Fig. 5-5b.
After the piston was pressed, the head of piston would depress the PDMS membrane in a concave
64

structure and push the liquid underneath. The membrane formed a closed region after the piston
was depressed, to prevent liquid leaking. However, the piston tended to make slight contact with
the PMMA via the top of notch after it was depressed, which did not provide sufficient area to keep
the piston steady, highlighted by the red circle as shown in Fig. 5-5b. When this occurred, the piston
would tilt causing the liquid to leak. Therefore, we modified the design by increasing the height of
the PMMA notch, enlarging the reservoir and increasing the piston contact area after it is depressed,
as shown in Fig. 5-5c. Also, a circular lid containing double-sided adhesive film was added to the
top of the piston to prevent liquid backflow after actuation. Based on this new design, the piston
remained oriented upright to prevent leaking. The adhesive film on the underside of the lid bonds
to the top surface of the chip after being depressed to prevent backflow.

Figure 5-5. Schematics of the three different designs of piston-based finger-powered pumps. Red
circles indicates defects during actuation. (a) The PDMS on the outer layer breaks due to the
friction force. (b) The depressed piston is tilting and results in leaking. (c) The optimal pump design.

Applying the piston pump design, a mock microfluidic cartridge was designed and filled
with colored dye for enhancing visualization. Red dye was added through the capillary tube, and
blue dye, DI water, and yellow dye were sequentially pumped into the sensing region from

65

reservoirs, as shown in Fig. 5-6. The performance of this design was superior to our prior designs
and did not suffer from poor sensor coverage or leakage. However, we observed another issue with
this design: bubble generation. Air inside the reservoirs formed bubbles in the microchannels after
actuation, resulting in channel blockage and poor sensor coverage. Although the piston-based pump
is functional, it suffers from poor reliability and requires a time-consuming fabrication process.

Figure 5-6. Microfluidic cartridge testing with colored dye. (a) Red dye was dispensed via capillary
tube, and blue dye (b) DI water (c) and yellow dye (d) were sequentially dispensed by depressing
the pistons.

5.5 Gasket-based piston pump

Figure 5-7. Schematic of the modified piston-based microfluidic pump before (left) and after (right)
actuation. Arrow indicates the rubber-based gasket to prevent liquid leaking.

66

To address the limitations of our piston-based pump described above, we modified the
design by incorporating a rubber gasket ring to secure the piston in the cartridge which provided a
leak-free seal. When the piston is depressed, it generates a uniform force on the liquid, which is
driven through the channel as shown in Fig. 5-7. Due to the simple fabrication process, this
technique provides further simplicity, reliability, and reproducibility for microfluidic device
generation. To evaluate the effectiveness of this design, mock microfluidic cartridges were
fabricated, filled with color dyes and tested.
To characterize the functionality of this pump, we performed several experiments to
evaluate the effectiveness of the pumping mechanism. First, we evaluated the minimum force to
overcome the friction between the rubber gasket and the piston to better understand the force
required to actuate the piston. We evaluated the force by placing different weights on top of the
piston and noted when the piston started moving downward. The measurements were performed
using five new chips and the average force was 1.35 N. However, we observed that the bottom of
the gasket became wet after the piston was depressed, which could influence the friction. Therefore,
we repeated the experiment with a pre-wetted gasket, and obtained an averaged force of 1.29 N.
This result indicates that the friction between the gasket and piston does not dependent on whether
the gasket is wet or dry.
To characterize the flow performance of this pump, flow rate measurements were
performed using microchannels having varying widths and a fixed pressing force of 1.84 N. As
shown in Fig. 5-8, the flow rate is proportional to the channel width where wider channels exhibited
larger flow rates. We briefly performed theoretical calculations of the flow rate inside a rectangular
microchannel using a simplified 2D model to compare with our experimental results. We first
estimated the Reynolds number (Re) using Eq. 5.1, where Îź is the fluid viscosity, Dh is the
hydraulic diameter of the channel, V is the flow velocity, and ρ is the fluid density. The hydraulic
diameter for a rectangular channel is defined as 4𝐴⁄𝑃, where A is the cross-section area, and P is
67

the wetted perimeter. The dimensions of the microchannels were measured using optical
microscopy, the flow velocities were measured using still frame images of video recordings of fluid
flow in the channel and we assumed the fluid properties of water at 20°C. For microchannels of
widths between 100 – 800 µm, the Re ranged from 0.1 to 0.5, which is fully laminar. By inserting
these values into Eq. 5.2, we obtain entrance lengths (Le) less than 10 Îźm. Since the length of the
microchannels is 15 cm, the entrance length is <0.01% of entire microchannel, and we can assume
that the flow is fully developed.

𝑅𝑒 =
𝐿𝑒
𝐷ℎ

𝜌𝑉𝐷ℎ

(5.1)

𝜇

≈ 0.06 𝑅𝑒

[115]

(5.2)

We also assumed that the flow inside the microchannel was steady with no-slip boundary
conditions at the channel walls and a uniform pressure gradient along the length of the channel.
Based on these assumptions, the volumetric flow rate inside a circular microchannel can be
described by the Hagen-Poiseuille law (Eq. 5.3), where Q is the volumetric flow rate, R is the radius,
L is the channel length and ΔP is the pressure drop across the entire channel.

𝑄=

𝜋𝑅 4 Δ𝑃
8𝜇 𝐿

(5.3)

Eq. 5.3 can be simplified as:
∆𝑃

𝑄=𝑅

[116]

𝐻

(5.4)

where RH is the hydraulic resistance, which is defined by the geometry of channel. The hydraulic
resistance for a rectangular channel can be calculated by the summation of Fourier series as shown
in Eq. 5.5 [117]
𝑅𝐻 =

12𝜇𝐿
ℎ 192

1

𝑤ℎ 3 (1− ( 5 ∑∞
𝑛=1(2𝑛−1)5 tanh(
𝑤 𝜋

68

(2𝑛−1)𝜋𝑤
)))
2ℎ

(5.5)

where w, h are the width and height of the channel, respectively. For channels with a high aspect
ratio (h/w< 1), which is typical of microfluidic systems, Eq. 5.5 can be reduced to the following:
𝑅𝐻 =

12𝜇𝐿
192ℎ
𝜋𝑤
[1 − 𝜋5 𝑤 tanh ( 2ℎ )]−1
𝑤ℎ 3

(5.6)

By combining Eq. 5.4 and 5.6, we can obtain an expression for Q inside a rectangular channel as a
function of ΔP and the channel geometry:
𝑄 = ∆𝑃 ×

𝑤ℎ 3
192ℎ
𝜋𝑤
[1 − 𝜋5 𝑤 tanh ( 2ℎ )]
12𝜇𝐿

(5.7)

To account for the friction between the piston and rubber gasket, 0.5 N was added to ΔP.
Comparison of the experimental and theoretical values obtained from Eq. 5.7 is shown in Fig. 5-8.

Flow Rate (ÎźL/s)

5

Theoretical

4

Experimental

3
2
1
0
0

200

400
600
Channel Width (Îźm)

800

Figure 5-8. Experimental vs. theoretical flow rates for rectangular microchannels 100 Âľm in height,
15 cm long with varying widths. Each data point is the mean Âą SD of five measurements using new
chips.

While the overall trends of the experimental and theoretical results are similar, the
deviation between the two sets of data is fairly significant, which we attribute to three sources of
experimental error. First, our theoretical analysis was based on a rectangular microchannel;

69

however, the sidewalls of the actual microchannels are slightly angled inward due to the laser
cutting process. Therefore, theoretical analysis based on a trapezoidal microchannel could provide
a more accurate comparison. Relatedly, the piston is not perfectly cylindrical due to the laser cutting
process and the area of the top surface is roughly 5% larger than the bottom surface. Therefore, as
the piston is depressed, the contact area with the gasket expands, resulting in a larger friction force.
This would cause the actual pressure exerted on the liquid to be less than we predicted, resulting in
a lower flow rate. Last, we observed that not all of the liquid is pushed into the microchannel when
the piston is pressed. There is a small gap between piston and reservoir which becomes filled with
the liquid. Thus, the actual pressure drop will be slightly less than we predicted.

5

Flow Rate (ÎźL/s)

4
3
2
1
0
5

10
15
Pressing Force (N)

20

Figure 5-9. Flow rate vs. pressing force for rectangular microchannels 100 Âľm high, 400 Âľm wide
and 10 cm long. Each bar represents the mean Âą SD of five separate measurements.

Next, we evaluated the effect of the pressing force on the flow rate. This experiment was
carried out using microchannels 400 Îźm wide, 100 Îźm high and 10 cm long filled with colored dye,
applying different weights to initiate the flow. As shown in Fig. 5-9, larger pressing forces
generated faster flow rates which were proportional to the pressure drop.

70

Moreover, we carried out an experiment to measure the pressing force required to
completely depress the piston with microchannels of varying widths filled will colored dye. As
shown in Fig. 5-10, the pressing force is inversely proportional to the channel width since the fluidic
resistance is a function of the channel size [118]. Specifically, larger pressing forces are required
to pump fluids inside smaller channels. The smallest channel that we tested (100 Âľm) required ~18
N of force to generate flow at 0.8 ÂľL/s, which is well within the capability of an average human
[119].

20
17.5

Force (N)

15
12.5
10
7.5
5
0

500

1,000
1,500
Channel Width (Îźm)

2,000

Figure 5-10. Pressing force required to pump water in rectangular microchannels 100 Âľm high,
10 cm long and of varying widths. Each data point represents the mean Âą SD of three separate
measurements.

Based on these results, the flow rate of this pump is influenced by the pressing force, which
can vary from test to test or person to person. Therefore, we investigated whether the flow rate
(pressing force) had a significant impact on the washing efficiency. To evaluate the influence of
the pressing force on the washing efficacy, we fabricated a microfluidic device (Fig. 5-11a)
incorporating a finger-powered pump and a reservoir pre-loaded with fluorescence dye and PBS,
respectively, and applied varying pressing forces (2, 5, 10, and 20 N) to depress the piston.
71

Figure 5-11. (a) Schematic of microfluidic device for evaluating washing efficiency. Fluorescent
images of the reservoir before (b) and after (c) washing. Scale bars, 100 Âľm. (d) Washing efficiency
in rectangular microchannels at 100 Âľm high and 400 Âľm of varying weights. Each bar represents
the mean Âą SD of three separate measurements.

The amount of residual dye was quantified by measuring the fluorescence signal in the
reservoir following washing via the finger-powered pump. The intensity of the fluorescence signals
were quantified using NIS-Element software. Measurements of the background signal (IB),
fluorescence intensity before washing (Ibefore), and after washing (Iafter) were performed using new
microfluidic chips. The background signal was measured at a region on the chip not containing any
fluorescence dye. The washing efficiency was calculated using the following equation:
72

𝑊𝑎𝑠ℎ𝑖𝑛𝑔 𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 =

𝐼𝑎𝑓𝑡𝑒𝑟 −𝐼𝐵
𝐼𝑏𝑒𝑓𝑜𝑟𝑒 −𝐼𝐵

× 100 %

(5.8)

The results of this experiment are shown in Figs. 5-11b-d. Fluorescence images of the
reservoir containing the fluorescence dye before and after washing are shown in Fig. 5-11b and c,
respectively. The washing efficiency for different pressing forces is presented in Fig. 5-11d, which
shows that there is not much improvement in the washing efficiency with pressing forces > 5 N.
Therefore, while larger pressing force can generate a faster flow rate, it doesn’t improve the
washing efficiency or influence the detection performance.
5.6 Summary
In this chapter, we have demonstrated a new finger-powered pump for transporting liquids
inside plastic microchannels. This pump consists of a rigid mechanical piston and a rubber gasket.
Depressing the piston generates a uniform pressure on the liquid inside the reservoir, which drives
it through the channel. The tight clearance between the piston and gasket prevents leaking and
backflow after activation. To characterize the effectiveness of this pumping mechanism for
transporting liquids inside microchannels, we measured the pressing force required for varying
channel widths. These results revealed that the pressing force is inversely proportional to the
channel width where larger forces are required to drive liquids in smaller channels. The smallest
channel that was tested (100 Âľm) requires ~18 N of force to generate flow at 0.8 ÂľL/s, which is
well within the capability of the average human. Additionally, we observed that the rate at which
the piston was depressed had a negligible impact on the flow performance. These results
demonstrate that our finger-powered pumping mechanism is a simple and effective method for
generating driving liquids in microfluidic devices.

73

Chapter 6. Microfluidic cartridge development
6.1 Motivation
In Chapters 4 and 5, we presented the development of an electrochemical assay for PfHRP2
detection and a finger-powered pump. In this chapter, we combine these technologies to develop a
finger-powered microfluidic electrochemical assay for rapid measurements of PfHRP2 in whole
blood samples. Two generations of prototypes were created: the first was based on the enzymatic
detection scheme (as described in Section 4.2) and the second was based on a nonenzymatic
detection scheme (as described in Section 4.3).
6.2 Materials and methods
6.2.1 Microfluidic cartridge design
The microfluidic cartridge consists of three layers of PMMA. The microfluidic network
was designed using AutoCAD software and etched into the PMMA using a CO2 laser cutter
(Universal Laser Systems, Scottsdale, AZ).
6.2.2 Microfluidic cartridge fabrication
The bottom layer consists of circular reservoirs (7 mm in diameter and 1.5 mm deep) for
storing liquid reagents. The electrochemical sensor, consisting of a working electrode (WE),
reference electrode (RE) and counter electrode (CE), was deposited on the back side of middle
layer through a shadow mask, and chromium/gold were sequentially evaporated with thicknesses
of 15 nm and 60 nm, respectively using a thermal evaporator (Jefferson Hills, PA).
1.5 mm-thick PMMA and rubber gaskets were obtained from McMaster-Carr (Elmhurst,
IL). Prior to fabrication, PMMA was rinsed with 2-propanol, deionized (DI) water and dried with
purified N2 gas. Capillary tubes were purchased from Sigma (St. Louis, MO) and double-sided
adhesive film was purchased from Adhesive Research (Glen Rock, PA). The middle and top layers
incorporate through-holes for the pistons, capillary tube inlet and air ports. Rubber gaskets were
74

installed between the middle and top layer, and all three layers were bonded together using doublesided adhesive. The pistons were laser cut from 3-mm-thick PMMA and inserted in each hole after
the reagents were dispensed in the reservoirs.
Chemical and protein preparation were carried out as described in Chapter 4, and the
fabrication of the finger-powered pump was carried out as described in Chapter 5. Microcontroller
LaunchPad and data converter chip were purchased from Texas Instruments (Dallas, TX).
Amplifiers and channel multiplexers were purchased from Analog Devices (Norwood, MA).
Discrete components and miscellaneous parts were purchased from Digi-Key (Thirf River Falls,
MN). For our prototype device, we used a Nexus 5 smartphone running on Android 5.0.2 OS.
6.2.3 Amperometric measurements
Amperometric measurements and signal processing were carried out by a custom detection
circuit, named aMEASURE [113] (Fig. 6-1), which was developed by Sina Parsnejad in Dr.
Andrew Mason’s lab at MSU.

Figure 6-1. Photograph of our mobile phone biosensing platform. The microfluidic chip (a) directly
plugs into aMEASURE (b), which is connected to the phone via the microUSB port using a USB
OTG adapter.

75

This circuit is comprised of a custom analog readout board and a microcontroller, which
provides a physical/electrical interface between microfluidic chip and the mobile phone. The circuit
can support a current signal from 0.94 nA to 2.4 mA, and automatically adapts to response current
levels. The circuit consumes very low power (90 mA) and can be powered for up to 7 hr using the
mobile phone battery. The microfluidic cartridge (Fig. 6-1a) plugs into aMEASURE via an
edgeboard connector. Step-by-step instructions for each measurement and test results are displayed
on the phone though a custom Android app, developed by Linlin Tu in Dr. Guoliang Xing's lab at
MSU.
6.3 First generation prototype device
6.3.1 Cartridge design
As shown in Fig. 6-2, our microfluidic cartridge consists of four components: reagent
storage reservoirs, finger-actuated pistons, microchannel network, and an electrochemical sensor.
The pistons are secured in place via rubber gaskets, which effectively seal off the reagents from the
environment. Due to the mild hydrophobicity of native PMMA, the liquid reagents remain
contained in the reservoirs until piston actuation. Two small air ports are placed downstream of the
reservoirs to allow air exhaustion during actuation, thereby preventing bubble generation inside the
fluidic network. The arrangement of the sample inlet, microchannels, and reagent reservoirs were
designed to facilitate the assay protocol and enhance the detection performance. The capillary tube
inlet is connected to the sensing region via a separate channel to minimize diffusion between the
sample and reagents. Separate channels were employed for the reporter solution (R1) and substrate
(R3) to prevent them from mixing, which would result in a high background signal. However,
merging the microchannels for the PBS wash buffer (R2) and substrate (R3) did not have a
significant impact on the background signal and was adopted to minimize cartridge size. Based on
these design strategies, our microfluidic cartridge does not require any additional fluidic
components, such as valves or diodes.

76

Figure 6-2. Exploded view of microfluidic cartridge (left). Photograph of the assembled
microfluidic cartridge with relevant regions labeled. Reservoirs R1, R2 and R3 are designed to
preload with a reporter solution, wash buffer and 4AP/H2O2 substrate, respectively. The
dimensions of the cartridge are 40 mm × 30 mm.

To simplify the testing process, all the sample processing steps are integrated on a
microfluidic cartridge. On the middle layer of the cartridge, we immobilized capture antibody on
gold sensor and applied blocker to prevent a false-positive signal from non-specific binding. Next,
we bonded the three PMMA pieces together using adhesive films. Color dyes were used to visualize
fluidic transport within the cartridge, as shown in Fig. 6-3.

Figure 6-3. Microfluidic cartridge testing with color dye. Red dye was added by capillary tube, and
blue dye, DI water and yellow dye were sequentially dispensed by depressing the pistons. (a) The
sample is dispensed from capillary tube and flows toward the sensor region. (b) Detection antibody
solution flows to the sensor by pressing piston R1. (c) Buffer washes away unbound detection
antibodies by pressing piston R2 (d) Substrate flows to the sensor region by pressing piston R3.

77

The blue dye represents the detection antibody stored in R1, water in R2 represents the washing
buffer to wash out the residue left on the sensor region, and the green dye in R3 represents the
substrate. The results of this experiment indicate that all three reagents can be sequentially
transported to the sensing region with negligible cross-contamination.
6.3.2 Device testing
To evaluate the performance of our prototype device, measurements were performed for
PfHRP2 detection, assay specificity and sensitivity. For assay specificity measurements, samples
were spiked with three different antigens (PfLDH, DEN-2 and PfHRP2) at 10 Âľg/mL in PBS. For
assay sensitivity measurements, human plasma samples were spiked with PfHRP2 from 0.1 Âľ/mL
to 20 Âľg/mL. 80 ÂľL of PBS containing 10 Âľg/mL of anti-PfHRP2 IgG, PBS and 4AP/H2O2 substrate
were dispensed into reservoirs R1, R2 an R3, respectively, using a pipette. To initiate the assay, 20
ÂľL of sample was dispensed into the cartridge using a capillary tube plunger. Next, a 400 mV
electrical field was applied to the sensor for 3 min to facilitate protein recognition and binding. The
value for voltage setting and incubation times are already characterized and addressed in Chapter
4. Px-conjugated detection antibody was delivered to the sensor by depressing piston R1. A second
electrical field (200 mV) was applied for 1.5 min, followed by a rinsing step to remove unbound
antibodies by pressing piston R2. Piston R3 was pressed to deliver the substrate solution, followed
by amperometry measurement at a bias potential of -150 mV for 1 min. The entire detection process
was completed in ≤ 6 min. Each measurement was performed at room temperature using a new
microfluidic cartridge.
We evaluated the specificity of our PfHRP2 assay by performing measurements using
irrelevant targets including dengue virus type 2 (DEN-2) and PfLDH, which is another malaria
biomarker. The measurements were taken for each sample at 10 Îźg/mL in PBS. As shown in Fig.
6-4, the sample containing PfHRP2 generated a SBR of ~16, whereas the irrelevant targets
generated low SBRs similar to PBS sample (blank control). These results suggest that the capture
78

and detection antibodies used in this assay specifically bind to only PfHRP2 thereby minimizing
the likelihood of false positive measurements.

20

SBR

15

10

5

0
PBS

DEN-2

PfLDH

PfHRP2

Figure 6-4. Specificity of PfHRP2 assay. Measurements were performed using DEN-2, PfLDH and
PfHRP2 in PBS at a concentration of 10 Îźg/mL, and PBS as a blank control. Each bar represents
of the mean Âą SD of three separate measurements.

Next, experiments were performed to measure PfHRP2 in human plasma samples at
different concentration from 0.1 Îźg/mL to 20 Îźg/mL. As shown in Fig. 6-5, amperometric signals
were clearly distinguishable at all of the tested concentrations with relatively smooth response
profiles and minimal noise. In addition, this assay can detect a wide range of concentrations (0.1
Âľg/mL - 20 Âľg/mL), encompassing the clinical levels found in patients with P. falciparum infection
[50], without sample dilution or concentration.

79

Figure 6-5. Amperometric signals of PfHRP2 diluted in human plasma. Measurements were
performed at a bias potential of -150 mV.

6.3.3 Summary
For the first generation device, a wide detection range of 0.1 Îźg/mL to 20 Îźg/mL was
achieved; however, the linearity was limited to 8 Îźg/mL. Liquid reagents are stored within the
cartridge, thereby simplifying the testing process. Although all the sample processing steps are
integrated in the cartridge, each measurement requires multiple steps, which can cause
inconsistencies in the test results due to human error. Therefore, a simple cartridge design requiring
a fewer number of steps is desired. In addition, the stability of lyophilized detection antibody
complex, dried 4AP and H2O2 is questionable. H2O2 tends to decompose slowly under room
temperature, however UV light catalyze the reaction. Additionally, H2O2 is a temperature sensitive
chemical, which does not allow any freeze-drying or special treatment. It’s difficult to dehydrate
H2O2 under room temperature. Therefore, we are not able to store H2O2 under room temperature
for a long time; alternative approaches for substrate selection and bioassay development are desired.

80

6.4 Second generation prototype device
We had a chance to demonstrate our first prototype and performed field-testing at the
College of Medicine at the University of Malawi in March, 2017. We received comments and
suggestions from the medical technicians there. We also noticed few disadvantages of first
generation microfluidic cartridge and devoted time to develop a second generation device. With
the redesigned device, it should provide more accurate results, simplify the processing steps and
offer an extended shelf life at room temperature.
Also, the first generation of our electrochemical instrument was quite bulky, (2 in. × 3 in.
× 1 in.). It is necessary to reduce the size of the instrument and make it portable and convenient for
POC application. With the development of an electrical interface, the second generation of
aMEASURE2 was developed, which is only 1 in. × 2 in. and consumes 250 mW (average) of power.
The second-generation prototype incorporates a smaller detection circuit, and the microfluidic
cartridge is connected to the circuit using a single-sided adapter, similar as the first generation
device.
6.4.1 Cartridge design
For the first generation of microfluidic cartridge, the fabrication process suffers from poor
alignment and results in an inconsistent device. Especially for the bonding of the top and middle
layers, poor alignment results in a deviated gasket position which affects the piston movement. The
consistency and accuracy of the alignment process affect the device quality and subsequent
measurement reliability. To improve the alignment accuracy, an alignment device was developed,
which is attached to the bottom of the cartridge, as shown in Fig. 6-6. The plastic holder was made
by 3D printing and consists of four, 1-mm-diameter extruded pillars located at the four corners.
Using this device, the lateral resolution of the alignment process is increased to 0.5 mm with
improved consistency.

81

Figure 6-6. Schematic of the alignment device with a lateral resolution of 0.5 mm.

For the second generation device, a second electrochemical sensor was added for a negative
control measurement. This negative control sensor does not contain PfHRP2-specific antibodies
and therefore generates a background signal. In the future, it is possible to modify this sensor to
detect another biomarker, which would enable multiplexed measurements for improved diagnostic
information. The microfluidic network was redesigned based on the updated detection scheme and
second sensor as shown in Fig. 6-7. This design incorporates dried reagent reservoirs and only 1
piston for the buffer. The layout of sample inlet, microchannels, sensors and dry reagents was
redesigned according to the nonenzymatic detection scheme.

Figure 6-7. Exploded view of microfluidic cartridge (left). Photograph of the assembled
microfluidic cartridge (right). Three reservoirs are reduced to one reservoir. The dimensions of
the cartridge are 30 mm × 25 mm.
82

As shown in Fig. 6-8a, dried Ru(NH3)63+ and MB were pre-stored in the reservoirs
respectively. Separate reservoirs isolated two chemicals before reconstitution. Also, the
microchannel network splits equally and contain serpentine regions to improve reconstitution. The
gold sensors are prepared using the same procedure as described in Section 6.2.2. Then, we
combined the target and detection antibody, adding the steps together. The dried detection antibody
is pre-stored in a reservoir designated for blood sample inlet. After the sample is added, the target
will mix the detection antibody and generate the binding based on the epitope recognition. Before
the mixture flow to the sensor region, we specifically incorporated a serpentine microchannel
before flowing to the sensor area as shown in Fig. 6-8b. With the extra serpentine area, there is
more contact area and enough time for improving the mixing and binding of the detection
antibody/target.
Once the detection antibody, target, and capture antibody are bound to the sensor surface,
a substrate solution is added for generating an electrochemical reaction. Unlike H2O2, Ru(NH3)63+
and MB are both stable at room temperature in dried form. As shown in Fig. 6-8a, we added 5 ÎźL
of Ru(NH3)63+ and MB in the reservoir, and let both chemicals dry in a low humidity chamber for
one day. Once dry, both chemicals are stable without requiring refrigeration. For reconstituting
Ru(NH3)63+ and MB on the cartridge, 150 ÎźL of PBS is added to the reservoir under the piston.
Once the piston is depressed, the liquid flows towards both regions and reconstitutes Ru(NH 3)63+
and MB, respectively. Followed by a merging region to mix Ru(NH3)63+ and MB, and flow to the
sensor region. As shown in Fig. 6-8c, compared with previous generation, there is less blood residue
remained on the sensor area due to the modified sensor region. Also, the liquid fills with all the
areas on the microfluidic cartridge; there is no observable dead volume; and a higher reconstitution
rate is developed for all the dried reagents on the cartridge.

83

Figure 6-8. (a) Schematic of updated microfluidic cartridge design. (b) Microfluidic cartridge
testing with color dye. Blood is added by capillary tube from inlet which contains detection
antibody. (c) The blood sample is pumped toward the outlet, and the mixture of MB and Ru(NH3)63+
flows to the sensor region.

6.4.2 Device testing
The second generation mobile phone biosensing system is shown in Fig. 6-9. The updated
microfluidic cartridge plugs into aMEASURE2 circuit board via an adapted connector. Step by step
instructions for each step in the detection process are displayed on the mobile phone through a
custom Android app interface.

84

Figure 6-9. Photo of the second generation mobile phone biosensing platform. The size of sensing
circuit is minimized to 2 in. × 1 in.

In order to develop a room-temperature stable microfluidic cartridge, dried reagents are
required, including the detection antibody, Ru(NH3)63+ and MB. Therefore, we optimized the
concentration of three reagents, respectively, and determined the optimal concentrations for each
reagent is 100 Îźg/mL, 1mM, and 0.5 mM, respectively. To maintain the activity and formation of
protein in a dry condition, the detection antibody is mixed with a 30% stabilizer solution to preserve
its activity. After the optimization of the incubation steps, the reagent condition, and applied
potential, a batch of microfluidic cartridges were prepared for device testing. A calibration curve
(Fig. 6-10) was generated by performing measurements of human blood samples spiked with
PfHRP2 from 500 ng/mL to 10,000 ng/mL. Although the detection range is narrower than our first
generation of microfluidic cartridge, it exhibits higher linearity (R2 = 0.971) which improves the
overall accuracy of this assay.

85

-700
-600

Current (nA)

-500

R² = 0.971

-400
-300
-200
-100
0
0

5,000
10,000
PfHRP2 Concentration (ng/mL)

Figure 6-10. The calibration curve generated from our POC system using spiked blood samples.
Each data point represents the mean Âą SD of three separate measurements obtained using new
cartridges.

One of advantages of the second generation microfluidic cartridge is the robustness of the
assay scheme. For this scheme, the cartridge can be stored under room temperature without
refrigeration due to the stability of Ru(NH3)63+ and MB. We then investigated the stability of the
microfluidic cartridge performance for a longer storage period. We pre-loaded 5 ÂľL of detection
antibodies, Ru(NH3)63+ and MB in the cartridge, and stored them in a dehumidified chamber, where
humidity is controlled under 30%. We performed measurements using microfluidic cartridge stored
for 1 week, 2 weeks, and 1 month to investigate the cartridge stability. For each period, we tested
with 1 Âľg/mL PfHRP2 spiked in human blood using three cartridges and the results are shown in
Fig. 6-11. The signals for the target and background remained stable for up to 2 weeks. However,
the target signal increased at 1 month storage due to a higher electron transfer rate occurring on the
surface. We suspect that the detection antibody does not exhibit the same activity after one month
storage, therefore there is less protein layer formed on the electrode, resulting in a higher current

86

signal. Additionally, the background signal dropped ~50 to 100 nA, suggesting that the substrate
activity in the cartridge slightly decreased as well. Although the SBR is decreased after 1 month of
storage, the detection signal can still be distinguished from the background.

-1,200

Background

Signal

Current (nA)

-1,000
-800
-600
-400
-200
0
0 week

1 week
2 weeks
Storage Duration

1 month

Figure 6-11. Detection current measurement for evaluating microfluidic cartridge stability at 1
Îźg/mL of PfHRP2 in human blood for 1 month. Each bar represents the mean Âą SD of three
separate measurements obtained using new cartridges.

6.4.3 Device validation
For the validation of our second generation mobile phone biosensor device, measurements
were performed using the mobile phone biosensor and a commercial PfHRP2 ELISA test.
Cartridges were prepared for evaluating samples from 1 Âľg/mL to 20 Âľg/mL using undiluted blood
samples by using capillary tubes pre-loaded with 20 ÂľL of PBS as a dilution buffer. Therefore, the
detection range of sample is expanded from 0.5 µg/mL – 10 µg/mL toward 1 µg/mL - 20 µg/mL.
Measurements obtained from the ELISA test (CellLab, Australia) were taken as the gold standard.
Analytical performance of our device was performed by testing 9 blood samples using both assays.

87

Phone Concentration Results (ng/mL)

25,000
20,000
R² = 0.983
15,000
10,000
5,000
0
0

5,000
10,000
15,000
20,000
ELISA Concentration Results (ng/mL)

25,000

Figure 6-12. Comparison of PfHRP2 measurements of spiked blood samples using ELISA and the
mobile phone device.

The result is shown in Fig. 6-12, our proposed device has a similar performance compared
with ELISA for measurements of PfHRP2 in spiked blood samples exhibiting good linearity (R2
correlation coefficient = 0.984). However, there is around 1,000 ng/mL concentration difference
for each measurement. Although the value on our device is not exactly same as ELISA, the signal
trends of the 9 samples are similar as ELISA. These preliminary results indicate that our POC test
can measure PfHRP2 in spiked whole blood samples with similar accuracy as ELISA.
6.5 Summary
In this chapter, we developed two generation prototypes of a finger-powered microfluidic
cartridge for rapid electrochemical measurements of PfHRP2 using a mobile phone biosensor
platform. For the first generation device, reagents are stored on-chip and fluid transport is
accomplished via a unique piston-based pumping mechanism. This electricity-free actuation
mechanism is coupled with a high sensitivity, high selectivity electrochemical assay which offers
rapid measurements and a wide detection range. For proof-of-concept, this platform was used for
88

quantitative measurements of PfHRP2 which could be detected from 0.1 µg/mL – 20 µg/mL in
spiked blood samples. Each measurement is completed in ≤ 6 min which is 3× faster than our
previously reported mobile phone electrochemical biosensor. However, the liquid reagents are
degraded on the microfluidic cartridge, the alternative approaches for storing reagent on cartridge
is necessary.
For the second generation device, we modified the design of the microfluidic cartridge to
incorporate a second sensor and employed a nonenzymatic detection scheme. By using a stabilizer,
reagents are dried on chip and can remain stable for up to a month at room temperature with
minimal signal loss. In addition, the testing procedure is simplified from three steps to one step.
For proof of concept, this platform was used for quantifying PfHRP2 in spiked blood samples with
high linearity from 0.5 µg/mL – 10 µg/mL. Each measurement is completed in ≤ 5 min which is
faster than our previously developed biosensor in Section 6.3. The validation of this device was
carried out by performing measurements of PfHRP2 in spiked blood samples and comparing the
results with a PfHRP2 ELISA test, which showed similar analytical performance.

89

Chapter 7. Conclusions
7.1 Summary of achievements and contributions
In this dissertation, we investigated various strategies towards the development of plastic
microfluidic devices for rapid electrochemical measurements of protein biomarkers. First,
UV/ozone (UVO) surface treatment was investigated for exploring wetting stability for 16 weeks
on three plastics and we noticed the storage condition has a significant impact on the surface
stability. Furthermore, we then explored two approaches for fabricating plastic microfluidic devices.
We demonstrated for the first time the fabrication of plastic microchannels using 3D printed metal
molds. Through optimization of the powder composition and processing parameters, we were able
to generate plastic microfluidic chips with similar flowing performance as device made from CNC
milled. However, this technique is replaced by CO2 ablation fabrication method due to the limited
resolution. In addition, PMMA was selected as the substrate material among three materials for the
microfluidic cartridge due to its stable chemical properties, and compatibility with CO2 ablation.
Two electrochemical assays for PfHRP2 with corresponding microfluidic cartridges were
developed, one based on an enzymatic detection scheme and another based on a nonenzymatic
detection scheme. Moreover, towards an electricity-free microfluidic assay, we developed a simple,
straightforward and valveless finger-powered pump, which was integrated with two
electrochemical assays. By utilizing a nonenzymatic MB-based detection scheme, the stability of
microfluidic cartridge is improved with a measurement time ≤ 5 min, which is 3× faster than the
previously reported mobile phone electrochemical biosensors [41,49]. A detection limit of 1 Îźg/mL
is achieved for this device. Although the sensitivity is lower than previously reported mobile phone
devices [41,48,49], this device exhibits a much higher linear detection range 1 Îźg/mL to 20 Îźg/mL,
which is required for the diagnosis of cerebral malaria. Furthermore, this device uses whole blood
samples, which circumvents complicated processing steps, such as centrifugation and dilution.

90

Owing to its portability, simplicity, and good analytical performance, this platform has the potential
to improve POC testing worldwide, particularly in remote and resource-limited settings.
7.2 Future work
The second generation microfluidic cartridge exhibits excellent stability at room
temperature for up to 1 month. By utilizing alternative methods to store dried reagents, it may be
possible to extend the shelf life of the cartridge. The current generation device employs laser
ablation and adhesive bonding which is simple, scalable, and reproducible. However, this approach
results in slight residue on the adhesive film and PMMA surfaces. Therefore, modifying the
fabrication process to minimize residue generation could improve the detection sensitivity and the
limit of detection. In the current prototype device, two sensors are employed for the detection signal
and a negative control. By incorporating additional sensors and modifying the microfluidic network,
it may be possible to perform multiplexed measurements of multiple biomarkers using a single
cartridge. This would provide more thorough diagnostic information and improve subsequent
treatment and patient outcomes.

91

BIBLIOGRAPHY

92

BIBLIOGRAPHY

[1]

B. Srinivasan and S. Tung, “Development and Applications of Portable Biosensors,” Jala,
vol. 20, no. 4, pp. 365–389, Aug. 2015.

[2]

P. Lems-Van Kan, H. W. Verspaget, and A. S. Peña, “ELISA assay for quantitative
measurement of human immunoglobulins IgA, IgG, and IgM in nanograms,” J. Immunol.
Methods, vol. 57, no. 1, pp. 51–57, Feb. 1983.

[3]

D. Grieshaber, R. MacKenzie, J. Vörös, and E. Reimhult, “Electrochemical Biosensors Sensor Principles and Architectures,” Sensors, vol. 8, no. 3, pp. 1400–1458, Mar. 2008.

[4]

J. P. Chambers, B. P. Arulanandam, L. L. Matta, A. Weis, and J. J. Valdes, “Biosensor
recognition elements,” Curr. Issues Mol. Biol., vol. 10, no. 1–2, pp. 1–12, 2008.

[5]

D. Mark, S. Haeberle, G. Roth, F. von Stetten, and R. Zengerle, “Microfluidic lab-on-a-chip
platforms: requirements, characteristics and applications,” Chem. Soc. Rev., vol. 39, no. 3, pp.
1153–1182, Feb. 2010.

[6]

G. A. Posthuma-Trumpie, J. Korf, and A. van Amerongen, “Lateral flow (immuno)assay: its
strengths, weaknesses, opportunities and threats. A literature survey,” Anal. Bioanal. Chem.,
vol. 393, no. 2, pp. 569–582, Jan. 2009.

[7]

K. M. Koczula and A. Gallotta, “Lateral flow assays,” Essays Biochem., vol. 60, no. 1, pp.
111–120, Jun. 2016.

[8]

I. Diaconu, C. Cristea, V. Harceaga, G. Marrazza, I. Berindan-Neagoe, and R. Sandulescu,
“Electrochemical immunosensors in breast and ovarian cancer,” Clin. Chim. Acta, vol. 425,
pp. 128–138, Oct. 2013.

[9]

D. Meadows, “Recent developments with biosensing technology and applications in the
pharmaceutical industry,” Adv. Drug Deliv. Rev., vol. 21, no. 3, pp. 179–189, Oct. 1996.

[10] E. B. Bahadir and M. K. Sezginturk, “Applications of electrochemical immunosensors for
early clinical diagnostics,” Talanta, vol. 132, pp. 162–174, Jan. 2015.
[11] A. Sassolas, L. J. Blum, and B. D. Leca-Bouvier, “Immobilization strategies to develop
enzymatic biosensors,” Biotechnol. Adv., vol. 30, no. 3, pp. 489–511, May 2012.
[12] J. Newman and K. E. Thomas-Alyea, Electrochemical Systems. John Wiley & Sons, 2012.
[13] M. Culebras, C. M. Gómez, and A. Cantarero, “Review on Polymers for Thermoelectric
Applications,” Materials, vol. 7, no. 9, pp. 6701–6732, Sep. 2014.
[14] J. Wang, “Electrochemical biosensors: Towards point-of-care cancer diagnostics,” Biosens.
Bioelectron., vol. 21, no. 10, pp. 1887–1892, Apr. 2006.
[15] X. Liu, T.-Y. Lin, and P. B. Lillehoj, “Smartphones for cell and biomolecular detection,” Ann.
Biomed. Eng., vol. 42, no. 11, pp. 2205–2217, Nov. 2014.
93

[16] L. C. Clark, “MONITOR AND CONTROL OF BLOOD AND TISSUE OXYGEN
TENSIONS.,” Trans.-Am. Soc. Fro Artif. Intern. Organs, vol. 2, no. 1, pp. 41–48, Apr. 1956.
[17] M. Liu, R. Liu, and W. Chen, “Graphene wrapped Cu2O nanocubes: Non-enzymatic
electrochemical sensors for the detection of glucose and hydrogen peroxide with enhanced
stability,” Biosens. Bioelectron., vol. 45, pp. 206–212, Jul. 2013.
[18] N. Ruecha, R. Rangkupan, N. Rodthongkum, and O. Chailapakul, “Novel paper-based
cholesterol biosensor using graphene/polyvinylpyrrolidone/polyaniline nanocomposite,”
Biosens. Bioelectron., vol. 52, pp. 13–19, Feb. 2014.
[19] N. J. Ronkainen-Matsuno, J. H. Thomas, H. B. Halsall, and W. R. Heineman,
“Electrochemical immunoassay moving into the fast lane,” TrAC Trends Anal. Chem., vol.
21, no. 4, pp. 213–225, Apr. 2002.
[20] A. Hemben, J. Ashley, and I. E. Tothill, “Development of an Immunosensor for PfHRP 2 as
a Biomarker for Malaria Detection,” Biosensors, vol. 7, no. 3, p. 28, Jul. 2017.
[21] W. S. Bauer et al., “Rapid concentration and elution of malarial antigen histidine-rich protein
II using solid phase Zn(II) resin in a simple flow-through pipette tip format,” Biomicrofluidics,
vol. 11, no. 3, p. 34115, May 2017.
[22] J. M. van Emon, Immunoassay and Other Bioanalytical Techniques. CRC Press, 2016.
[23] E. V. Gogol, G. A. Evtugyn, J.-L. Marty, H. C. Budnikov, and V. G. Winter, “Amperometric
biosensors based on nafion coated screen-printed electrodes for the determination of
cholinesterase inhibitors,” Talanta, vol. 53, no. 2, pp. 379–389, Nov. 2000.
[24] S. Laschi, I. Palchetti, and M. Mascini, “Gold-based screen-printed sensor for detection of
trace lead,” Sens. Actuators B Chem., vol. 114, no. 1, pp. 460–465, Mar. 2006.
[25] C. R. Ispas, G. Crivat, and S. Andreescu, “Review: Recent Developments in Enzyme-Based
Biosensors for Biomedical Analysis,” Anal. Lett., vol. 45, no. 2–3, pp. 168–186, Jan. 2012.
[26] J. S. Swensen et al., “Continuous, Real-Time Monitoring of Cocaine in Undiluted Blood
Serum via a Microfluidic, Electrochemical Aptamer-Based Sensor,” J. Am. Chem. Soc., vol.
131, no. 12, pp. 4262–4266, Apr. 2009.
[27] M. Pumera, A. Merkoçi, and S. Alegret, “New materials for electrochemical sensing VII.
Microfluidic chip platforms,” TrAC Trends Anal. Chem., vol. 25, no. 3, pp. 219–235, Mar.
2006.
[28] P. N. Nge, C. I. Rogers, and A. T. Woolley, “Advances in Microfluidic Materials, Functions,
Integration, and Applications,” Chem. Rev., vol. 113, no. 4, pp. 2550–2583, Apr. 2013.
[29] D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, “Rapid Prototyping
of Microfluidic Systems in Poly(dimethylsiloxane),” Anal. Chem., vol. 70, no. 23, pp. 4974–
4984, Dec. 1998.
[30] W. Dungchai, O. Chailapakul, and C. S. Henry, “Electrochemical Detection for Paper-Based
Microfluidics,” Anal. Chem., vol. 81, no. 14, pp. 5821–5826, Jul. 2009.
94

[31] Z. Nie et al., “Electrochemical sensing in paper-based microfluidic devices,” Lab Chip, vol.
10, no. 4, pp. 477–483, Feb. 2010.
[32] M. Reches, K. A. Mirica, R. Dasgupta, M. D. Dickey, M. J. Butte, and G. M. Whitesides,
“Thread as a Matrix for Biomedical Assays,” ACS Appl. Mater. Interfaces, vol. 2, no. 6, pp.
1722–1728, Jun. 2010.
[33] D. J. Laser and J. G. Santiago, “A review of micropumps,” J. Micromechanics
Microengineering, vol. 14, no. 6, p. R35, 2004.
[34] P. Woias, “Micropumps—past, progress and future prospects,” Sens. Actuators B Chem., vol.
105, no. 1, pp. 28–38, Feb. 2005.
[35] C. Jönsson et al., “Silane–dextran chemistry on lateral flow polymer chips for immunoassays,”
Lab Chip, vol. 8, no. 7, pp. 1191–1197, Jun. 2008.
[36] I. Beaulieu, M. Geissler, and J. Mauzeroll, “Oxygen plasma treatment of polystyrene and
Zeonor: substrates for adhesion of patterned cells,” Langmuir ACS J. Surf. Colloids, vol. 25,
no. 12, pp. 7169–7176, Jun. 2009.
[37] P. M. van Midwoud, A. Janse, M. T. Merema, G. M. M. Groothuis, and E. Verpoorte,
“Comparison of Biocompatibility and Adsorption Properties of Different Plastics for
Advanced Microfluidic Cell and Tissue Culture Models,” Anal. Chem., vol. 84, no. 9, pp.
3938–3944, May 2012.
[38] J. Castillo-LeĂłn and W. E. Svendsen, Lab-on-a-Chip Devices and Micro-Total Analysis
Systems: A Practical Guide. Springer, 2014.
[39] D. S. Kim, S. H. Lee, C. H. Ahn, J. Y. Lee, and T. H. Kwon, “Disposable integrated
microfluidic biochip for blood typing by plastic microinjection moulding,” Lab Chip, vol. 6,
no. 6, pp. 794–802, May 2006.
[40] K. Iwai, K. C. Shih, X. Lin, T. A. Brubaker, R. D. Sochol, and L. Lin, “Finger-powered
microfluidic systems using multilayer soft lithography and injection molding processes,” Lab.
Chip, vol. 14, no. 19, pp. 3790–3799, Oct. 2014.
[41] T. Laksanasopin et al., “A smartphone dongle for diagnosis of infectious diseases at the point
of care,” Sci. Transl. Med., vol. 7, no. 273, p. 273re1-273re1, Feb. 2015.
[42] T.-S. Leu and P.-Y. Chang, “Pressure barrier of capillary stop valves in micro sample
separators,” Sens. Actuators Phys., vol. 115, no. 2–3, pp. 508–515, Sep. 2004.
[43] H. Andersson, W. van der Wijngaart, P. Griss, F. Niklaus, and G. Stemme, “Hydrophobic
valves of plasma deposited octafluorocyclobutane in DRIE channels,” Sens. Actuators B
Chem., vol. 75, no. 1–2, pp. 136–141, Apr. 2001.
[44] M. L. Adams, M. L. Johnston, A. Scherer, and S. R. Quake, “Polydimethylsiloxane based
microfluidic diode,” J. Micromechanics Microengineering, vol. 15, no. 8, p. 1517, 2005.
[45] D. Quesada-González and A. Merkoçi, “Mobile phone–based biosensing: an emerging
‘diagnostic and communication’ technology,” Biosens. Bioelectron.
95

[46] J. Jiang et al., “Smartphone based portable bacteria pre-concentrating microfluidic sensor and
impedance sensing system,” Sens. Actuators B Chem., vol. 193, pp. 653–659, Mar. 2014.
[47] A. F. Coskun, R. Nagi, K. Sadeghi, S. Phillips, and A. Ozcan, “Albumin testing in urine using
a smart-phone,” Lab. Chip, vol. 13, no. 21, pp. 4231–4238, Nov. 2013.
[48] C. C. Stemple, S. V. Angus, T. S. Park, and J.-Y. Yoon, “Smartphone-based optofluidic labon-a-chip for detecting pathogens from blood,” J. Lab. Autom., vol. 19, no. 1, pp. 35–41, Feb.
2014.
[49] P. B. Lillehoj, M.-C. Huang, N. Truong, and C.-M. Ho, “Rapid electrochemical detection on
a mobile phone,” Lab Chip, vol. 13, no. 15, pp. 2950–2955, Jul. 2013.
[50] K. B. Seydel et al., “Plasma concentrations of parasite histidine-rich protein 2 distinguish
between retinopathy-positive and retinopathy-negative cerebral malaria in Malawian
children,” J. Infect. Dis., vol. 206, no. 3, pp. 309–318, Aug. 2012.
[51] A. Piruska et al., “The autofluorescence of plastic materials and chips measured under laser
irradiation,” Lab. Chip, vol. 5, no. 12, pp. 1348–1354, Dec. 2005.
[52] S. H. Tan, N.-T. Nguyen, Y. C. Chua, and T. G. Kang, “Oxygen plasma treatment for reducing
hydrophobicity of a sealed polydimethylsiloxane microchannel,” Biomicrofluidics, vol. 4, no.
3, p. 32204, Sep. 2010.
[53] A. Bhattacharyya and C. M. Klapperich, “Mechanical and chemical analysis of plasma and
ultraviolet–ozone surface treatments for thermal bonding of polymeric microfluidic devices,”
Lab Chip, vol. 7, no. 7, pp. 876–882, Jun. 2007.
[54] C. W. Tsao, L. Hromada, J. Liu, P. Kumar, and D. L. DeVoe, “Low temperature bonding of
PMMA and COC microfluidic substrates using UV/ozone surface treatment,” Lab Chip, vol.
7, no. 4, pp. 499–505, Mar. 2007.
[55] M. Malmsten, “Ellipsometry studies of the effects of surface hydrophobicity on protein
adsorption,” Colloids Surf. B Biointerfaces, vol. 3, no. 5, pp. 297–308, Jan. 1995.
[56] M. Rabe, D. Verdes, and S. Seeger, “Understanding protein adsorption phenomena at solid
surfaces,” Adv. Colloid Interface Sci., vol. 162, no. 1–2, pp. 87–106, Feb. 2011.
[57] Y. Arima and H. Iwata, “Effect of wettability and surface functional groups on protein
adsorption and cell adhesion using well-defined mixed self-assembled monolayers,”
Biomaterials, vol. 28, no. 20, pp. 3074–3082, Jul. 2007.
[58] S. Pasche, J. Vörös, H. J. Griesser, N. D. Spencer, and M. Textor, “Effects of Ionic Strength
and Surface Charge on Protein Adsorption at PEGylated Surfaces,” J. Phys. Chem. B, vol.
109, no. 37, pp. 17545–17552, Sep. 2005.
[59] H. Becker and L. E. Locascio, “Polymer microfluidic devices,” Talanta, vol. 56, no. 2, pp.
267–287, Feb. 2002.

96

[60] D. J. Guckenberger, T. E. de Groot, A. M. D. Wan, D. J. Beebe, and E. W. K. Young,
“Micromilling: a method for ultra-rapid prototyping of plastic microfluidic devices,” Lab
Chip, vol. 15, no. 11, pp. 2364–2378, May 2015.
[61] M. E. Wilson et al., “Fabrication of circular microfluidic channels by combining mechanical
micromilling and soft lithography,” Lab. Chip, vol. 11, no. 8, pp. 1550–1555, Apr. 2011.
[62] S.-C. Wang, C.-Y. Lee, and H.-P. Chen, “Thermoplastic microchannel fabrication using
carbon dioxide laser ablation,” J. Chromatogr. A, vol. 1111, no. 2, pp. 252–257, Apr. 2006.
[63] H. Klank, J. P. Kutter, and O. Geschke, “CO(2)-laser micromachining and back-end
processing for rapid production of PMMA-based microfluidic systems,” Lab. Chip, vol. 2,
no. 4, pp. 242–246, Nov. 2002.
[64] A. K. Au, W. Lee, and A. Folch, “Mail-order microfluidics: evaluation of stereolithography
for the production of microfluidic devices,” Lab. Chip, vol. 14, no. 7, pp. 1294–1301, Apr.
2014.
[65] P. J. Kitson, M. H. Rosnes, V. Sans, V. Dragone, and L. Cronin, “Configurable 3D-Printed
millifluidic and microfluidic ‘lab on a chip’ reactionware devices,” Lab Chip, vol. 12, no. 18,
pp. 3267–3271, Aug. 2012.
[66] Y. Zhao and T. Cui, “Fabrication of high-aspect-ratio polymer-based electrostatic comb
drives using the hot embossing technique,” J. Micromechanics Microengineering, vol. 13, no.
3, p. 430, 2003.
[67] M. B. Esch, S. Kapur, G. Irizarry, and V. Genova, “Influence of master fabrication techniques
on the characteristics of embossed microfluidic channels,” Lab Chip, vol. 3, no. 2, pp. 121–
127, May 2003.
[68] S. Miserere, G. Mottet, V. Taniga, S. Descroix, J.-L. Viovy, and L. Malaquin, “Fabrication
of thermoplastics chips through lamination based techniques,” Lab Chip, vol. 12, no. 10, pp.
1849–1856, Apr. 2012.
[69] L. J. Kricka, P. Fortina, N. J. Panaro, P. Wilding, G. Alonso-Amigo, and H. Becker,
“Fabrication of plastic microchips by hot embossing,” Lab. Chip, vol. 2, no. 1, pp. 1–4, Feb.
2002.
[70] J. Greener et al., “Rapid, cost-efficient fabrication of microfluidic reactors in thermoplastic
polymers by combining photolithography and hot embossing,” Lab. Chip, vol. 10, no. 4, pp.
522–524, Feb. 2010.
[71] R. Novak, N. Ranu, and R. A. Mathies, “Rapid fabrication of nickel molds for prototyping
embossed plastic microfluidic devices,” Lab Chip, vol. 13, no. 8, pp. 1468–1471, Mar. 2013.
[72] Q. Xia and S. Y. Chou, “Fabrication of sub-25 nm diameter pillar nanoimprint molds with
smooth sidewalls using self-perfection by liquefaction and reactive ion etching,”
Nanotechnology, vol. 19, no. 45, p. 455301, Nov. 2008.

97

[73] V. N. Goral, Y.-C. Hsieh, O. N. Petzold, R. A. Faris, and P. K. Yuen, “Hot embossing of
plastic microfluidic devices using poly(dimethylsiloxane) molds,” J. Micromechanics
Microengineering, vol. 21, no. 1, p. 17002, 2011.
[74] G. S. Fiorini, G. D. M. Jeffries, D. S. W. Lim, C. L. Kuyper, and D. T. Chiu, “Fabrication of
thermoset polyester microfluidic devices and embossing masters using rapid prototyped
polydimethylsiloxane molds,” Lab Chip, vol. 3, no. 3, pp. 158–163, Aug. 2003.
[75] T. Koerner, L. Brown, R. Xie, and R. D. Oleschuk, “Epoxy resins as stamps for hot embossing
of microstructures and microfluidic channels,” Sens. Actuators B Chem., vol. 107, no. 2, pp.
632–639, Jun. 2005.
[76] R. K. Jena, C. Y. Yue, and K. X. Yun, “Effect of a CNT based composite micromold on the
replication fidelity during the microfabrication of polymeric microfluidic devices,” RSC Adv.,
vol. 4, no. 24, pp. 12448–12456, Feb. 2014.
[77] E. W. K. Young et al., “Rapid Prototyping of Arrayed Microfluidic Systems in Polystyrene
for Cell-Based Assays,” Anal. Chem., vol. 83, no. 4, pp. 1408–1417, Feb. 2011.
[78] J. A. Olortegui-Yume and P. Y. Kwon, “Crater wear patterns analysis on multi-layer coated
carbides using the wavelet transform,” Wear, vol. 268, no. 3, pp. 493–504, Feb. 2010.
[79] M. Vaezi and C. K. Chua, “Effects of layer thickness and binder saturation level parameters
on 3D printing process,” Int. J. Adv. Manuf. Technol., vol. 53, no. 1–4, pp. 275–284, Mar.
2011.
[80] T. Do et al., “Improving Structural Integrity with Boron-based Additives for 3D Printed 420
Stainless Steel,” Procedia Manuf., vol. 1, pp. 263–272, Jan. 2015.
[81] S. Sandron et al., “3D printed metal columns for capillary liquid chromatography,” Analyst,
vol. 139, no. 24, pp. 6343–6347, Nov. 2014.
[82] H. Becker and U. Heim, “Hot embossing as a method for the fabrication of polymer high
aspect ratio structures,” Sens. Actuators Phys., vol. 83, no. 1–3, pp. 130–135, May 2000.
[83] L. Brown, T. Koerner, J. H. Horton, and R. D. Oleschuk, “Fabrication and characterization
of poly(methylmethacrylate) microfluidic devices bonded using surface modifications and
solvents,” Lab. Chip, vol. 6, no. 1, pp. 66–73, Jan. 2006.
[84] Y. Sun, Y. C. Kwok, and N.-T. Nguyen, “Low-pressure, high-temperature thermal bonding
of polymeric microfluidic devices and their applications for electrophoretic separation,” J.
Micromechanics Microengineering, vol. 16, no. 8, p. 1681, 2006.
[85] E. M. Hamad et al., “Inkjet printing of UV-curable adhesive and dielectric inks for
microfluidic devices,” Lab Chip, vol. 16, no. 1, pp. 70–74, Dec. 2015.
[86] J.-Y. Cheng, C.-W. Wei, K.-H. Hsu, and T.-H. Young, “Direct-write laser micromachining
and universal surface modification of PMMA for device development,” Sens. Actuators B
Chem., vol. 99, no. 1, pp. 186–196, Apr. 2004.

98

[87] C.-W. Tsao and D. L. DeVoe, “Bonding of thermoplastic polymer microfluidics,” Microfluid.
Nanofluidics, vol. 6, no. 1, pp. 1–16, Jan. 2009.
[88] J. P. Gosling, “A decade of development in immunoassay methodology,” Clin. Chem., vol.
36, no. 8 Pt 1, pp. 1408–1427, Aug. 1990.
[89] D. Kim and A. E. Herr, “Protein immobilization techniques for microfluidic assays,”
Biomicrofluidics, vol. 7, no. 4, Jul. 2013.
[90] S. Sharma, H. Byrne, and R. J. O’Kennedy, “Antibodies and antibody-derived analytical
biosensors,” Essays Biochem., vol. 60, no. 1, pp. 9–18, Jun. 2016.
[91] F. Darain, M. A. Wahab, and S. C. Tjin, “Surface Activation of Poly(Methyl methacrylate)
by Plasma Treatment: Stable Antibody Immobilization for Microfluidic Enzyme-Linked
Immunosorbent Assay,” Anal. Lett., vol. 45, no. 17, pp. 2569–2579, Nov. 2012.
[92] T. T. Le, C. P. Wilde, N. Grossman, and A. E. G. Cass, “A simple method for controlled
immobilization of proteins on modified SAMs,” Phys. Chem. Chem. Phys. PCCP, vol. 13,
no. 12, pp. 5271–5278, Mar. 2011.
[93] A. Ahmad and E. Moore, “Electrochemical immunosensor modified with self-assembled
monolayer of 11-mercaptoundecanoic acid on gold electrodes for detection of benzo[a]pyrene
in water,” Analyst, vol. 137, no. 24, pp. 5839–5844, 2012.
[94] N. R. Mohamad, N. H. C. Marzuki, N. A. Buang, F. Huyop, and R. A. Wahab, “An overview
of technologies for immobilization of enzymes and surface analysis techniques for
immobilized enzymes,” Biotechnol. Biotechnol. Equip., vol. 29, no. 2, pp. 205–220, Mar.
2015.
[95] Q. Yu, Q. Wang, B. Li, Q. Lin, and Y. Duan, “Technological Development of Antibody
Immobilization for Optical Immunoassays: Progress and Prospects,” Crit. Rev. Anal. Chem.,
vol. 45, no. 1, pp. 62–75, Jan. 2015.
[96] J. Wang, G. Liu, M. H. Engelhard, and Y. Lin, “Sensitive Immunoassay of a Biomarker
Tumor Necrosis Factor-α Based on Poly(guanine)-Functionalized Silica Nanoparticle Label,”
Anal. Chem., vol. 78, no. 19, pp. 6974–6979, Oct. 2006.
[97] K. Tanaka and H. Imagawa, “Rapid and reagent-saving immunoassay using innovative
stirring actions of magnetic beads in microreactors in the sequential injection mode,” Talanta,
vol. 68, no. 2, pp. 437–441, Dec. 2005.
[98] V. N. Morozov and T. Y. Morozova, “Electrophoresis-Assisted Active Immunoassay,” Anal.
Chem., vol. 75, no. 24, pp. 6813–6819, Dec. 2003.
[99] W. J, Y. Y, Y. F, and J. H, “Electric field-driven strategy for multiplexed detection of protein
biomarkers using a disposable reagentless electrochemical immunosensor array.,” Anal.
Chem., vol. 80, no. 15, pp. 6072–6077, 2008 2008.
[100] D. Du, J. Wang, D. Lu, A. Dohnalkova, and Y. Lin, “Multiplexed Electrochemical
Immunoassay of Phosphorylated Proteins Based on Enzyme-Functionalized Gold Nanorod

99

Labels and Electric Field-Driven Acceleration,” Anal. Chem., vol. 83, no. 17, pp. 6580–6585,
Sep. 2011.
[101] J. Tate and G. Ward, “Interferences in Immunoassay,” Clin. Biochem. Rev., vol. 25, no. 2, pp.
105–120, May 2004.
[102] A. I. Alayash, R. P. Patel, and R. E. Cashon, “Redox Reactions of Hemoglobin and
Myoglobin: Biological and Toxicological Implications,” Antioxid. Redox Signal., vol. 3, no.
2, pp. 313–327, Apr. 2001.
[103] “Pierce Traut’s Reagent (2-iminothiolane) - Thermo Fisher Scientific.” [Online]. Available:
https://www.thermofisher.com/order/catalog/product/26101. [Accessed: 02-Jan-2018].
[104] P. Fanjul-Bolado, M. B. González-García, and A. Costa-García, “Amperometric detection in
TMB/HRP-based assays,” Anal. Bioanal. Chem., vol. 382, no. 2, pp. 297–302, May 2005.
[105] P. A. Rogers, W. M. Chilian, I. N. Bratz, R. M. Bryan, and G. M. Dick, “H2O2 activates
redox- and 4-aminopyridine-sensitive Kv channels in coronary vascular smooth muscle,” Am.
J. Physiol. Heart Circ. Physiol., vol. 292, no. 3, pp. H1404-1411, Mar. 2007.
[106] D. B. Rye, C. B. Saper, and B. H. Wainer, “Stabilization of the tetramethylbenzidine (TMB)
reaction product: application for retrograde and anterograde tracing, and combination with
immunohistochemistry.,” J. Histochem. Cytochem., vol. 32, no. 11, pp. 1145–1153, Nov.
1984.
[107] H. J. Kang, M. A. Aziz, B. Jeon, K. Jo, and H. Yang, “Strategy for Low Background-Current
Levels in the Electrochemical Biosensors Using Horse-Radish Peroxidase Labels,”
Electroanalysis, vol. 21, no. 24, pp. 2647–2652, Dec. 2009.
[108] M. Zhang and W. Gorski, “Electrochemical Sensing Platform Based on the Carbon
Nanotubes/Redox Mediators-Biopolymer System,” J. Am. Chem. Soc., vol. 127, no. 7, pp.
2058–2059, Feb. 2005.
[109] J.-Z. Xu, J.-J. Zhu, Q. Wu, Z. Hu, and H.-Y. Chen, “An Amperometric Biosensor Based on
the Coimmobilization of Horseradish Peroxidase and Methylene Blue on a Carbon Nanotubes
Modified Electrode,” Electroanalysis, vol. 15, no. 3, pp. 219–224, Mar. 2003.
[110] H. Yao, N. Li, S. Xu, J.-Z. Xu, J.-J. Zhu, and H.-Y. Chen, “Electrochemical study of a new
methylene blue/silicon oxide nanocomposition mediator and its application for stable
biosensor of hydrogen peroxide,” Biosens. Bioelectron., vol. 21, no. 2, pp. 372–377, Aug.
2005.
[111] J. Wu, Y. Yan, F. Yan, and H. Ju, “Electric Field-Driven Strategy for Multiplexed Detection
of Protein Biomarkers Using a Disposable Reagentless Electrochemical Immunosensor
Array,” Anal. Chem., vol. 80, no. 15, pp. 6072–6077, Aug. 2008.
[112] K. Iwai, K. C. Shih, X. Lin, T. A. Brubaker, R. D. Sochol, and L. Lin, “Finger-powered
microfluidic systems using multilayer soft lithography and injection molding processes,” Lab
Chip, vol. 14, no. 19, pp. 3790–3799, Aug. 2014.

100

[113] C. Peng and Y. S. Ju, “Finger-Powered Electro-Digital-Microfluidics,” Methods Mol. Biol.
Clifton NJ, vol. 1572, pp. 293–311, 2017.
[114] K. Xu, M. R. Begley, and J. P. Landers, “Simultaneous metering and dispensing of multiple
reagents on a passively controlled microdevice solely by finger pressing,” Lab Chip, vol. 15,
no. 3, pp. 867–876, Jan. 2015.
[115] L. S. Han, “Hydrodynamic Entrance Lengths for Incompressible Laminar Flow in
Rectangular Ducts,” J. Appl. Mech., vol. 27, no. 3, pp. 403–409, Sep. 1960.
[116] K. W. Oh, K. Lee, B. Ahn, and E. P. Furlani, “Design of pressure-driven microfluidic
networks using electric circuit analogy,” Lab Chip, vol. 12, no. 3, pp. 515–545, 2012.
[117] R. J. Cornish and M. Sc, “Flow in a pipe of rectangular cross-section,” Proc R Soc Lond A,
vol. 120, no. 786, pp. 691–700, Oct. 1928.
[118] B. J. Kirby, Micro- and Nanoscale Fluid Mechanics: Transport in Microfluidic Devices.
Cambridge University Press, 2010.
[119] C. A. Crosby and M. A. Wehbé, “Hand strength: Normative values,” J. Hand Surg., vol. 19,
no. 4, pp. 665–670, Jul. 1994.
[120] J. H. Sung and M. L. Shuler, “Prevention of air bubble formation in a microfluidic perfusion
cell culture system using a microscale bubble trap,” Biomed. Microdevices, vol. 11, no. 4, pp.
731–738, Aug. 2009.

101