Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2007

Development of a Digital Microfluidic Lab-on-a-Chip for Automated Immunoassay withMagnetically Responsive BeadsRamakrishna Sista

Follow this and additional works at the FSU Digital Library. For more information, please contact lib-ir@fsu.edu

THE FLORIDA STATE UNIVERSITY

FAMU-FSU COLLEGE OF ENGINEERING

DEVELOPMENT OF A DIGITAL MICROFLUIDIC LAB-ON-A-CHIP FOR AUTOMATED IMMUNOASSAY WITH

MAGNETICALLY RESPONSIVE BEADS

By

RAMAKRISHNA SISTA

A Dissertation submitted to the Department of Chemical Engineering

In partial fulfillment of the Requirements for the degree of

Doctor of Philosophy

Degree Awarded: Spring Semester, 2007

The members of the Committee approve the dissertation of Ramakrishna Sista defended on

29th

March, 2007.

Srinivas Palanki

Professor Directing Dissertation

Jim Zheng

Outside Committee Member

John C.Telotte

Committee member

Ching Jen Chen

Committee member

Approved:

Bruce Locke, Chair, Department of Chemical Engineering

Ching Jen Chen, Dean, College of Engineering

The Office of Graduate Studies has verified and approved the above named committee

members

ii

iii

To My Parents…

iv

ACKNOWLEDGEMENTS

I would like to express my deep sense of gratitude to my advisor Dr. Srinivas Palanki for

his guidance, help and encouragement throughout the research. I would also like to thank

the other members of my committee, Dr. John Telotte, Dr. Ching Jen Chen and Dr. Chifu

Wu for their valuable suggestions and encouragement. I would also like to thank Dr.

Yousef Haik and Dr. Jhunu Chatterjee for introducing me to the project and providing me

with the magnetic beads at Florida State University. I am forever indebted to Dr. Vamsee

Pamula and Dr. Vijay Srinivasan for introducing me to the concept of droplet based

transport using electrowetting and for their guidance through out the project. I extend my

heartfelt gratitude to Dr. Allen Echkhardt for his guidance and support through out the

research. His enthusiasm for my results was one of the main driving forces for my research.

I would also like to express my special thanks to Dr. Phil Paik and Dr. Dwayne Allen for

assisting me in setting up experiments and for their insightful discussions. I am grateful to

each and everyone at Advanced Liquid Logic, Inc. for making my dissertation project a

pleasant experience. Last but not the least, I would like to thank my family and friends for

their continuous support and encouragement through out the process.

v

TABLE OF CONTENTS

List of Tables…………………………………………………… ix

List of Figures………………………………….......................... x

Abstract………………………………………………. ….......... xiii

1 Background and Motivation………………………………................... 1

1.1 Introduction……………………………………………………………. 1

1.2 Integrated microfluidic devices……………………………………. 2

1.3 Point-of-care testing (POCT)………………………………..……... 3

1.4 Scope and Objective…………………………………………............. 5

2 Microfluidics in Clinical diagnostics…………………………………. 6

2.1 Conventional clinical diagnostics…………………………………. 6

2.2 Description of an immunoassay……………………………………. 6

2.3 Commercial automated immunoassay analyzers……………….... 7

2.4 State-of-the-art Lab-on-a-chip prototypes………………………… 11

2.5 Commercial microfluidic devices for immunoassays on chip…. 12

2.6 Diagnosis of Insulin and IL-6……………………………………..... 15

2.6.1 Interleukin-6…………………………………………………… 15

2.6.2 Insulin………………………………………………………… 16

2.6.3 Current diagnostic techniques to detect Insulin and IL-6………. 16

2.7 Chapter Summary……………………………………………………. 17

3 Digital microfluidics………………………………………............................ 18

3.1 Droplet based microfluidics…………………………………………. 18

3.2 Surface tension driven flow.....……………………………………… 18

vi

3.3 Electrowetting based droplet actuation……………………..…….. 21

3.3.1 Theory of electrowetting………………………………………... 21

3.4 Electrowetting based lab-on-a-chip for point of care testing…. 24

3.4.1 Droplet dispensing……………………………………………… 24

3.4.2 Droplet transport…………………………………………… …... 25

3.4.3 Sample preparation and droplet dilution………………………… 25

3.4.4 Efficiency and speed of droplet mixing……………………........ 25

3.4.5 Biocompatibility of electrowetting platform……………………. 27

3.5 Chapter Summary……………………………………………………… 30

4 Lab-on-a-chip design, fabrication and testing………………………. 31

4.1 Lab-on-a-chip specifications……………………...…………………. 31

4.2 Fabrication of lab-on-a-chip and system assembly……………… 31

4.2.1 Chip fabrication……………………………………………… 31

4.2.2 Top plate fabrication…………………………………………..... 33

4.2.3 System assembly………………………………………………... 33

4.3 Detection Instrumentation………………………………………….... 35

4.4 Design of the reservoir and the droplet pathways………………. 39

4.5 Biocompatibility of the lab-on-a-chips…………………………..... 41

4.6 Protein fouling on chips…………………………………………….. 47

4.7 Material defects……………………………………………………….. 47

4.8 Chapter Summary…………………………………………………….. 49

5 On chip magnetic immunoassay……………………............................. 51

5.1 Description of an immunoassay…………………………………… 51

5.2 Super paramagnetic beads………………………..………………… 52

5.3 Parameters involved in washing of magnetic beads……...…..... 52

5.3.1 Buffer system…………………………………………………... 53

5.3.2 Magnetic field strength…………………………………………... 57

vii

5.3.3 Concentration of the bead suspension…………………………… 59

5.3.4 Position of the magnet…………………………………………… 59

5.4 Washing of the magnetic beads……………………………………… 70

5.4.1 Bead retention……………………………………………………. 70

5.4.2 Dilution efficiency………………………………………………... 72

5.5 Magnetic Immunoassay of on chip……………………..…………… 72

5.5.1 Experimental setup………………………………………………. 73

5.5.2 Chemiluminescence detection………………………………….... 75

5.5.3 Experimental protocol (on chip)…………………………………. 77

5.5.4 Experimental protocol (on bench)………………………………… 80

5.6 Insulin assay on serum…………………………. …………………….. 83

5.7 Magnetic immunoassay of Interleukin-6 (IL-6) on chip…….…….86

5.7.1 Immunoassay on digital microfluidic chip…………………………86

5.7.2 Immunoassay of IL-6 on bench…………………………………… 88

5.8 Immunoassay of IL-6 on serum………………….………………….... 90

5.9 Cross contamination of the analytes with magnetic beads……...…92

5.10 Chapter Summary…….…………………………………………..……. 93

6 Conclusions and Future work……………………………………………... 94

6.1 Conclusions……………………………………………………………… 94

6.1.1 Lab-on-a-chip fabrication and testing…………………………….. 94

6.1.2 Washing of magnetic beads……………………………………….. 95

6.1.3 Magnetic Immunoassay of Insulin and IL-6 on droplet

based lab-on-a-chip………………………………………………… 96

6.2 Future work………………………………………………………………. 97

6.2.1 Lab-on-a-chip architecture………………………………………... 97

6.2.2 Detection methodologies………………………………………..... 98

6.2.3 System integration………………………………………………… 98

viii

Appendix A: Analysis of Chemiluminescent data using enzyme kinetics................ 100

Appendix B: Image processing using Image J software……………........................ 102

References………………………………………………………………………… 106

Biographical sketch……………………………………………………………. 113

ix

LIST OF TABLES

1.1 Compounds analyzed in a clinical laboratory with times for monitoring………… 3

1.2 Comparison between laboratory based system and point-of-care system………... 4

5.1 Parameters to characterize magnetic bead attraction…………………………….. 53

5.2 Washing done on bench-scale equipment and on-chip…………………………… 75

5.3 Reagents in the Ultra-sensitive Insulin Immunoassay kit….……………………… 75

5.4 Reagents in the Access Immunoassay IL-6 kit…………………………………… 86

x

LIST OF FIGURES

2.1 Schematic representation of a magnetic immunoassay…………………………… 8

2.2 Abbott AxSYM automated immunoassay analyzer………………………………. 10

2.3 Access Immunoassay analyzer………………………………………………… …. 10

2.4 Gyros compact disc microfluidic chip…………………………………………….. 14

2.5 Micronics microfluidic card……………………………………………………….. 14

3.1 Cross section of equilibrium forces acting on one side of non-wetting droplet in contact with a horizontal surface…...…………………………………………..... 20 3.2 Cross section of a droplet resting on a surface having a gradient of surface energy… 20 3.3 The electrowetting effect……………………………………………………………. 23

3.4 Top and side views of a basic electrowetting effect………………………………… 23

3.5 Dispensing of KCl droplets from a reservoir……………………………………….. 26

3.6 Transport of droplet on a 2 phase bus……………………………………………….. 26

3.7 Frequency versus voltage curves for various physiological fluids………………….. 29

3.8 Transport of blood droplet on transparent ITO electrodes………………………….. 29

4.1 Lab-on-a-chip design to perform immunoassay involving magnetic beads………… 32

4.2 Fabrication process for droplet based lab-on-a-chip for immunoassays……………. 34

4.3 Working of a PMT based on external photoelectric effect………..………………… 36

4.4 Detection instrumentation for chemiluminescent detection in the droplet based lab- on-a-chip………………………………………………………………………………… 37 4.5 Circuit for chemiluminescent detection of the signal using a PMT …………….. 38

4.6 Scheme of dispensing a droplet from a reservoir on lab-on-a-chip made in printed circuit board…………………………………………………………………… 40 4.7 Transport of unit droplet containing magnetically responsive beads across five electrodes……………………………………………………………………………... 43 4.8 Dispensing and transport of serum on chip……………………………………….. 44

4.9 Asymmetric splitting of 0.01% Tween® 20 in PBS………………………………... 46

4.10 Symmetric splitting of water droplet…………………………………………… 46

4.11 Fouling of HRP enzyme on chip………………………………………………… 48

xi

4.12 Electrolysis in the reservoir……………………………………………………….. 50

4.13 Electrolysis at the point of splitting……………………………………………….. 50

5.1 BioMAG streptavidin coated beads with no surfactant in the supernatant (Image taken after 40 seconds)…..………………………………………………….. 55 5.2 BioMAG streptavidin coated beads with 0.005% w/w Tween 20 in the supernatant (Image taken after 40 seconds)………...…………………………. 55 5.3 BioMAG streptavidin coated beads with 0.01% w/w Tween 20 in the supernatant (Image taken after 40 seconds)………………………………………. 56 5.4 BioMag streptavidin beads in 0.01 % Tween® 20 attracted with a 0.5 Tesla magnet (5 lbs pull force) (Image after 40 seconds)..................…………… 58 5.5 BioMAG stretavidin beads in 0.01% Tween® 20 attracted with a 0.5 Tesla magnet (1.25 lbs pull force) (Image after 40 seconds)…………………... 58 5.6 BioMag Streptavidin beads (undiluted stock) in 0.01% Tween® 20 (Image after 40 seconds)…………………………………………………………… 60 5.7 BioMag Streptavidin beads (4 times diluted) in 0.01% Tween® 20 (Image after 40 seconds)………..………………………………………………… 60 5.8 One Tesla magnet placed underneath the electrode with the bead droplet………… 62

5.9 Simulation of magnetic field lines of a 1Tesla magnet placed underneath a magnetic bead droplet…………………………………………………………… 62

5.10 One Tesla magnets placed underneath and over the magnetic bead droplet……… 63

5.11 Simulation of magnetic field lines of 1 Tesla magnets placed underneath and over a magnetic bead droplet………….……………………………………. 63 5.12 One Tesla magnets placed on four sides of the magnetic bead droplet…………… 65

5.13 Simulation of the magnetic field lines of the quadrapole arrangement………… 65

5.14 Splitting mechanism to retain the magnetic beads and remove the supernatant…. . 67

5.15 Schematic representation of the splitting of the supernatant retaining the magnetic beads………………………………………………………………… 67 5.16 Loss of magnetic beads with the splitting occurring within the effect of the magnetic field……………………………………………………………………… 68 5.17 Splitting of the supernatant occurring away from the magnetic field…………….. 69

5.18 96 well plate comparing the washes of magnetic beads on bench and chip………. 74

xii

5.19 Comparison of washing on chip and on conventional bench-scale equipment….... 74

5.20 Chemiluminescence detection setup……………………………………………… 76

5.21 Step by step protocol of magnetic immunoassay on chip…………………………. 78

5.22 Kinetic curves of different concentrations of insulin in a magnetic immunoassay on chip…………………………………………………………… 79 5.23 Insulin standard curve on lab-on-a-chip……………………………………… 79

5.24 Kinetic curves of different concentrations of insulin in a magnetic immunoassay on bench………………...………………………………………… 81 5.25 Insulin standard curve on conventional bench-scale equipment..……………… 81

5.26 Comparison of Insulin standard curves on conventional bench-scale equipment and chip…………….………………………………………………………………….…. 82 5.27 Kinetic curves of magnetic immunoassay of serum for Insulin………………… 84

5.28 Kinetic curve of magnetic immunoassay on serum for Insulin on conventional bench-

scale equipment……………………………………………………………...………… 85

5.29 Kinetic curve of magnetic immunoassay on serum for Insulin on chip…………. 85

5.30 Kinetic curves of IL-6 immunoassay on lab-on-a-chip………………………… 88

5.31 Standard curve of IL-6 immunoassay on lab-on-a-chip…………………………. 88

5.32 Standard curve of IL-6 immunoassay on conventional bench-scale equipment.... 89

5.33 Comparison of IL-6 magnetic immunoassay on chip and bench……………… 89

5.34 Kinetic curve of magnetic immunoassay of serum for IL-6 on chip…………… 91

5.35 Kinetic curve of magnetic immunoassay of serum for IL-6 on conventional bench-

scale equipment…………………………………………………………..…………… 91

B.1 Analyzing particles using Image J………………………………………………... 102

B.2 Watershedding using Image J……………………………………………………... 103

B.3 Thresholded image of magnetic beads in 0.01% Tween 20 ……………………… 103

B.4 Thresholded image of magnetic beads in 0.005% Tween 20 …............................. 104

B.5 Thresholded image of magnetic beads in 0% Tween 20………………………….. 104

B.6 Thresholded image of undiluted stock of magnetic beads………………………… 105

B.7 Thresholded image of beads attracted with a magnet of high pull force (5 lbs)…... 105

.

xiii

ABSTRACT

The emerging paradigm of lab-on-a-chip powered by microfluidics is expected to

revolutionize miniaturization, automation and integration in the point-of-care centers which

require quick, efficient and reproducible results. Furthermore, high throughput requirement

from the life sciences laboratories have made the development of lab-on-a-chip a major

area of research over the past few years. Immunoassays, which are routinely used to

determine the concentrations of various analytes in life sciences laboratories, are one of the

most important laboratory tests that require efficient automation and integration. These are

different from other laboratory tests in the clinical laboratories such as colorimetric tests

because they involve formation of antibody-antigen complexes to generate a signal that can

be measured. The immunoassays also involve the application of magnetically responsive

beads to increase the surface area for the reactions thereby enhance the signal. Although

miniaturization was started in the early 1990’s there are no commercial devices that

perform immunoassays involving magnetic beads. None of the state-of-the art commercial

microfluidic technologies, which are based on continuous flow in etched microchannel,

have been able to fully deliver the promised benefits of microfluidics. This is primarily due

to their incompatability with common sample matrices and architectural inflexibility.

Furthermore, the transport of the magnetic beads in microchannels is a difficult task in

continuous mode of operation as magnetic susceptibility of the beads is rather weak and

because of demagnetization of the particles. In this thesis, a droplet-based microfluidic lab-

on-a-chip based on electrowetting actuation is developed to perform immunoassays using

magnetic beads on human physiological samples. Biocompatibility of the electrowetting

system is established by demonstrating repeatable and rapid transport of human

physiological fluids such as whole blood, serum, proteins such as bovine serum albumin,

antibodies for insulin and interleukin-6 and enzymatic reagents such as horseradish

peroxidase (HRP), alkaline phosphatase (ALP). Various magnetic configurations for

efficient attraction of the magnetic beads that assist in the washing are developed. Several

parameters involved in washing magnetic beads are studied and the buffer for resuspending

the beads, magnetic strength to attract the magnetic beads and concentration of beads to be

xiv

used were established. An efficient protocol for washing magnetic beads on chip was

developed with a bead retention efficiency of almost 100%. A complete magnetic

immunoassay is performed on chip for Insulin and Interleukin-6. The least concentration of

Insulin and IL-6 detectable on chip is 0.24 pg/µL and 4 fg/µL respectively. Standard curves

are developed for both the analytes over a range of concentrations. The repeatability of the

assays is established by performing the assay on different samples on different days and the

standard error is shown to be less than 3%. Magnetic immunoassays on the droplet based

lab-on-a-chip are also performed on serum for Insulin and IL-6 and it is shown that the

results are comparable to data obtained via conventional laboratory analysis. This work

represents the first demonstration of integrated and automated operation of a digital-

microfluidic lab-on-a-chip for immunoassays involving magnetically responsive beads on

clinically relevant sample matrices.

1

CHAPTER 1

BACKGROUND AND MOTIVATION

1.1 Introduction

Since clinical laboratories started using robotics in the early 1980s to manage their

thousands of samples everyday, automated machines are routinely used to manage

thousands of samples everyday [Chapman et al., 2003]. Automation of the clinical

laboratories can significantly increase throughput by rapidly performing processes right

from sample storage and retrieval to running assays [Gwynne et al., 2002]. This frees up

researchers from performing repetitive tasks and allows them to focus on monitoring and

analyzing the raw data produced. This can vastly reduce human intervention, which in turn

could improve the quality of the data produced. Apart from this, high levels of parallelism

and its associated throughput benefits are also enabled by automation of labs and

integration of processes. Automation and integration also enables analytical instruments to

be used for point-of-care-testing (POCT) which is becoming a major force in the future

evolution of health care [Bissell et al., 2002]. Point of care testing devices currently

account for almost one fourth of the world-wide market for clinical laboratory in vitro

diagnostic products and are projected to double over the next decade [Stephans et al.,

1999].

There are several reasons for turning from macroscale analytical procedures to microscale

analytical devices. Some of these reasons include low sample consumption, fast analysis

times and high throughput potentialities, the possibility of integration and automation, and

the desire for single-use devices associated with suppression of cross-contamination [Lion

et al., 2004].

The significant increase in the number of samples analyzed in clinical laboratories is

helping to drive laboratories into the automated world of industrial production lines. The

first use of automation and integration began with mechanized robots for fluid handling in

2

micro-titer plates. These systems were initially designed to handle simple and repetitive

tasks like liquid handling and pipetting but have constantly and steadily grown to perform

more complicated tasks based on the emerging laboratory needs. Current robots available

commercially can handle sample volumes of microliters to tens of nanoliters. However the

volume reproducibility reduces as the volume scales down to nanoliters (6% at 100 nL for

Caliper’s Sciclone ALH 3000 as compared to 2 % at 1 µL). Apart from this, the robotics

involved in handling such low volumes is prohibitively expensive and occupy large spaces.

The cost of these machines limits their usage only to large labs with high throughput

applications.

1.2 Integrated microfluidic devices: Lab-on-a-chip

Modern microfluidic devices can be traced back to development of the first

miniaturized device on silicon for gas chromatography at Stanford university [Terry et al.,

1979] and the ink jet printer at IBM [Petersen et al., 1979; Bassous et al., 1977]. Although

both these devices were quite remarkable, the new paradigm of integrated microfluidic

devices, which are also popularly referred to as Lab-on-a-chip (LoC) or Micro Total

Analysis Systems (µTAS) [Reyes et al., 2002], was developed only in the early 1990s by

Manz et al. [1990]. The basic functionalities of µTAS include sample preparation, sample

injection, fluid and particle handling, reaction and mixing, separation and detection

[Auroux et al., 2002]. All these operations are analogous to those performed using

laboratory scale equipment in a conventional laboratory. Since the advent of the LoC

[Erickson et al., 2004], the field has blossomed and branched off into many different areas

citing different applications of the same: biological and chemical analysis [Beebe et al.,

2002; Jakeway et al., 2000], [Chovan et al., 2002], point of care testing [Tudos et al.,

2001], clinical and forensic analysis [Verpoorte et al., 2002], molecular diagnostics [Huang

et al., 2002] and medical diagnostics [Vo-Dinh et al., 2000].

A microfluidic Lab-on-a-chip (LoC) can handle sub nanoliter volumes without loss in

reproducibility because of the nanometer range dimensions of the devices manufactured

using highly efficient lithographic processes. This also reduces the manufacturing costs

considerably [Kock et al., 2000]. Furthermore, LoC systems occupy significantly less

laboratory space as compared to conventional robotic systems. The applications of LoC

3

systems in the point of care testing (POCT) market accounts for almost one fourth of the

world’s market for in vitro diagnostics and is projected to double over the next decade

[Stephans et al., 1999].

1.3 Point-of-Care Testing (POCT)

Presently, disease prevention and treatment is based on measurement of certain chemical

parameters in physiological fluids like blood and urine. For all these chemical parameters,

samples are typically transported to a central laboratory for analysis and the results of the

assays or the clinical tests might take several hours or days. The table below shows the

examples of typical compounds along with the time scale for monitoring [Tudos et al.,

2001].

Table 1.1 Compounds analyzed in a clinical laboratory with times for monitoring

Seconds/minutes Hours Days

Oxygen Creatinine Iron

Carbon di oxide Bilirubin Albumin

Potassium Urea Globulin

Glucose Sodium Cholestrol

Lactate Chloride

Cortisol Triglyceride

Neurotransmitters

Apart from time, cost also plays a very important factor in the growth of point of care

testing. It was reported that POCT introduces almost 35 % savings per analysis and

additional savings on man power [Adams et al., 1995]. POCT eliminates the need for many

current activities involving in sending blood samples to a remote laboratory setting.

Furthermore, the time consuming steps of labeling, icing and bagging samples can be

eliminated and the potential for lost, broken, or clotted specimens are minimized [Fleisher

et al., 1993]. Elimination of such type of errors can result in a diminished need for repeated

4

sampling especially in the case of neonatal babies [Fini et al., 1994] and pediatric patients.

Table 1.2 gives a comparison between the point-of-care testing and the regular laboratory

based system [Tudos et al., 2001].

Table 1.2 Comparison between laboratory based system and a point-of-care system

Laboratory based system Point-of-care system

1. Test is ordered 1. Test is ordered

2. Test request is processed 2. Collection of blood sample

3. Collection of blood sample 3. Analysis of sample

4. Transportation of the sample 4. Review of results

5. Labeling and storage 5. Clinicians act on results

6. Centrifugation of the sample

7. Sorting to analyzers

8. Analysis of samples

9. Review of the results

10. Clinicians act on results

Hence there is a need for developing POCT devices for doctor’s offices and hospitals to

expedite the testing process, thereby reducing the hospital costs. According to Chan et al.

[1996], “The clinical laboratory of the future will be organized differently than today’s

laboratory: testing will be performed at various locations, both within and outside the

traditional hospital environment, laboratories will be operated with a higher degree of

automation, the job description of laboratory personnel will be changed to include more

information management responsibilities, and both laboratory costs and services will

become highly competitive”.

5

1.4 Scope and objective

The overall objective of this thesis is to design and build a nano-liter lab-on-a-chip

platform to perform an integrated and automated immunoassay involving magnetically

responsive beads. In Chapter 2, the applications of microfluidics in POCT are described

and a description of various commercial devices involving microfluidics is presented. In

Chapter 3, the concept of droplet-based transport is described and a brief literature survey

of the applications of droplet based transport in clinical chemistry and polymerase chain

reaction (PCR) is presented. In Chapter 4, the design and fabrication of lab-on-a-chip for

performing immunoassays involving magnetic beads is described. The testing of the

fabricated chip with the reagents involved in the immunoassay and the mobility of the

magnetic beads on a microfluidic lab-on-a-chip is also presented. Material issues for

fabrication of the chips are also presented in this chapter. In Chapter 5, several parameters

involved in washing of magnetic beads are described and optimum conditions to achieve

efficient washing of the magnetic beads are presented. The complete magnetic

immunoassays of insulin and interleukin-6 (IL-6) are described in this chapter. Conclusions

and possible future extensions of this work are presented in Chapter 6.

6

CHAPTER 2

MICROFLUIDICS IN CLINICAL DIAGNOSTICS

2.1 Conventional clinical diagnostics

Clinical diagnostics in humans are commonly performed on physiological fluids such as

blood, serum, plasma or urine. Other fluids, such as saliva [Streckfus et al., 2002], sweat

[Taylor et al., 1996], and cerebral spinal fluid, have also been used to detect specific

antigens. The number of analytes that can be assayed in a commercial high through-put

clinical diagnostic analyzer ranges from 60-130 [Aller et al., 2002]. Based on the biological

function these antigens/analytes can be categorized into amino acids, proteins [Lee et al.,

2001; Mouradian et al., 2002; Doherty et al, 2003], cytokines [Morse et al, 2004],

enzymes, carbohydrates, lipids, vitamins, electrolytes, blood gases, trace elements, drugs

[Huels et al, 2002], tumor markers and nucleic acids [Burtis et al, 1999]. Among the

various methods available, immunoassays are widely used in molecular biology providing

the molecular basis for many clinical diagnoses [Eteshola et al.., 2001; Bernard et al..,

2001].

2.2 Description of an immunoassay

An immunoassay is a biochemical test that measures the concentration of a protein or

hormone in a physiological liquid like blood, serum or urine [Gosling et al., 2005].

Immunoassays benefit from very high selectivity and affinity of antibody/antigen systems,

as well as from decades of immunoassay developments in diagnostics. The detection of

concentration of the antibody or the antigen can be achieved by variety of methods. Some

of the most common techniques are to label the antigen or antibody with an enzyme

(Enzyme linked immunoassay), radio isotopes (Radio immunoassay) or a fluorescent tag.

Other techniques include agglutination [Piras et al., 2005; Qiaojia et al., 1998],

nephelometry, turbidimetry [Delanghe et al., 1991 and western blot [Rodgers et al., 1984].

Immunoassays are classified as homogenous and heterogeneous assays. In a homogenous

immunoassay, also referred to as the competitive immunoassay, the antigen in the unknown

7

sample competes with the labeled antigen to bind to the antibodies. The amount of labeled

antigen to the antibodies is then measured by different detection techniques as mentioned

above. The response or signal measured is inversely proportional to the concentration of

the antigen in the unknown. This is because greater the response, the less antigen in the

unknown was able to “compete” with the labeled antigen for binding with the antibodies. In

a heterogeneous immunoassay, also referred to as the “sandwich assay,” a “sandwich”

complex is formed with an antigen coupled between a primary antibody and a secondary

antibody. The primary antibody is immobilized onto a solid surface which may be typically

a plate or the surface of a tube or a bead.

The secondary antibody is labeled with an enzyme which reacts with a substrate or

an introduced chemical reagent to give a relative indication of the concentration of the

antigen (the extent of the reaction is a relative measure of the concentration of the antigen).

The antigen of interest is captured between the two antibodies as shown in Figure 2.1

which can further be separated from the unreacted solution and analyzed. A heterogeneous

immunoassay includes an extra step of removing the excess or unbound antibody or antigen

from the reaction site, using a solid phase reagent usually which is the solid wall of a tube

or a plate or beads made of various materials. The immunoassay which utilizes

magnetically responsive beads or spheres as the solid phase is termed as “magnetic

immunoassay”.

2.3 Commercial automated immunoassay analyzers

Immunoassays are among the most sensitive and specific analytical immunological

methods used in clinical diagnostics. Among the various immunoassay methodologies, the

Enzyme linked Immunosorbent Assay (ELISA) is the most common because of its high

specificity and sensitivity. An immunoassay requires a large number of repetitive steps

resulting in long processing times and high labor costs.

8

Magnetic bead bound

antiibody

Antigen Enzyme bound

antibody

Key to Figure

Antibody

Antigen

Enzyme

label

Magnetic bead

Figure 2.1 Schematic representation of a magnetic immunoassay

9

Automation of such processes would significantly reduce the cost of the tests and the time

for result [Chan et al, 1996]. Access® Immunoassay analyzer from Beckmann Coulter Inc.,

Ciba Corning ACS:180 automated immunoassay system [Brugues et al, 1994], Tosoh AIA-

600 from GMI and Abbott AxSYMTM [Smith et al, 1993] from Abbott laboratories are the

most popular among the commercially available automated immunoassay analyzers. The

basic characteristics of these automated immuno-analyzers include repetitive pipetting of

reagents with the help of mechanized robots. Apart from this, these analyzers are also

equipped with integrated chemiluminescent/optical/fluorescent detection equipment. The

Access® automated analyzer has a throughput of 50-100 tests per hour with sample sizes

varying from 10-200µL and is equipped with chemiluminescent detection [Patterson et al.,

1994]. Though significant advances have been made in the automation of immunoassays,

these analyzers are prohibitively expensive ($50,000-$200,000) and are not affordable in a

low-throughput research setting or a point-of-care setting like a physician’s office. The

more affordable lower end systems with automated plate washers, incubators and

integrated optics for detection still require a skilled technician to perform several key steps

in an immunoassay such as preparing microtiter plates with antibodies and loading samples

onto the plates. This may result in human error due to repeated manual intervention and is a

major source of inter-assay and intra-assay variation. Furthermore these immunoassay

analyzers require significant laboratory space (Figures 2.2 and 2.3) which limits their usage

at the centers of point of care. The microfluidic (bio) analytical devices [Lion et al., 2004]

have received much attention over the past 15 years especially because their potential to

significantly reduce the processing time and cost for analysis in genomics and proteomics

research. The motivation for turning from macroscale analytical procedures to microscale

analytical devices includes low sample consumption, less waste production, fast analysis

times and high throughput potentialities, integration and automation potential and the

desire for single use devices associated with the suppression of cross-contamination

10

Figure 2.2 Abbott AxSYMTM automated immunoassay analyzer

Figure 2.3 Access® Immunoassay analyzer

11

2.4 State-of-the-art Lab-on-a-chip prototypes

The use of microfluidic lab-on-chip technology for clinical diagnostic applications has been

reviewed extensively by Verpoorte et al. (2002) and Tudos et al. (2001). The emerging

paradigm of a lab-on-a-chip powered by microfluidic systems [Reyes et al, 2002; Iossifidis

et al, 2002] favorably enables mass transport limited reactions such as ELISA, due to its

inherently small length scales (in the order of microns). Diffusion times are proportional to

the square of diffusion length and therefore reducing the length scale from 1 mm

(microliter) to 0.1 mm (nanoliters) reduces the incubation time of the analyte with the

antibodies by 2 orders of magnitude-from hours on the micro titer plate to minutes in the

microfluidic system. Microfluidic lab-on-a-chip device platforms also offer a high degree

of integration and automation at a fraction of cost of the robotic systems, significantly

reducing the technician costs and also minimizing human error. Miniaturization,

automation and faster detection also enables the use of microfluidic chips in a point-of-care

setting [Frost et a.l, 2003]. Currently almost all the microfluidic devices are based on

continuous fluid flow in permanent microchannels fabricated in glass, plastic or other

polymers. The most commonly used fluid actuation methods for pumping involve

electrokinetic phenomena using electroosmosis and electrophoresis for separation.

Electrokinetic methods scale favorably with miniaturization and do not require any

mechanical pumps or valves for fluidic control. Electrophoresis has been widely used in

clinical diagnostics applications for the separation of large class of analytes including

amino acids, DNA fragments, carbohydrates, proteins, drugs and vitamins [Burtis et al.,

1999; Thormann et al., 2001]. A general review of the microfluidic maniupulations and

their applications in electro-kinetically driven microfludic chips was done by Bruin et al.

(2000).

Though electrokinetic actuation methods are most commonly used for micro-

channel based microfluidics there is a trend towards use of alternative fluid actuation

techniques since many sample matrices are not directly compatible with electrokinetic

phenomena [Verpoorte et al., 2002]. This is due to their high sensitivity to liquid properties

such as pH and ionic strength. For example physiological fluids like blood and urine cannot

be pumped using the electroosmotic flow because of excessive Joule heating [Duffy et al.,

1999].

12

The second most common actuation strategy to move fluids is a positive

displacement method using syringe pumps [Leach et al., 2003; Moser et al., 1997; Hayashi

et al., 2003; Nakamura et al., 2001]. This technique has been combined with in-vivo micro

dialysis sampling for continuous on-line monitoring of small molecules such as glucose,

lactate, and chloride [Bohm et al, 2000; Dempsey et al, 1997]. However these continuous

microfluidic systems require complicated valving schemes if a higher degree of fluid

control is required. This also increases the degree of complexity in fabrication.

2.5 Commercial microfluidic devices for immunoassays on chip

Currently, development of integrated devices for immunoassays is significantly less

advanced than that for DNA analysis. Hence there are only a handful of devices available

in the market despite extensive research in microfluidic technology. Furthermore, most of

these devices do not actually have enough functionality and flexibility to be called a lab-

on-a-chip. A brief review of the commercially available microfluidic devices for clinical

diagnostics and immunoassays is presented in the following paragraphs.

Caliper’s next generation LabChip® 3000 [Anonymous et al., 2007a] drug

discovery system which uses mobility shift electrophoresis for separation of the product

and substrate and performs fluorogenic enzymatic assays. In fact it is the only commercial

instrument that uses electrophoresis for actuation. Other passive schemes using capillary

forces and external pressure driven devices are more common due to their compatibility

with a wider range of samples. The Gyrolab microlaboratory from Gyros [Anonymous et

al., 2007b] and LabCD integrated microfluidic system from Tecan [Anonymous et al.,

2007c] developed a bio-analytical system composed of a disposable microfluidic compact

device (CD) (Figure 2.4) that is intended for sandwich immunoassays and an instrument for

automatic processing of CDs (GyrolabTM workstation). The liquid movement and

localization are achieved by a combination of capillary action, centrifugal force and

hydrophobic barriers within the microstructure. Micronics’ ORCA microfluidics

[Anonymous et al., 2007d] platform uses various pressure driven microfluidic elements,

such as the diffusion based laminar flow diffusion based H-Filter® technology in a

disposable Active HTM card (Figure 2.5) used in medical diagnostics, biopharmaceutical

13

and other analytical applications. The BiositeTM Triage cardiac system [Anonymous et al.,

2007e]] measures cardiac markers in whole blood in a microcapillary based device for

point-of-care testing. Another popular point-of-care testing analyzer is the iSTAT

[Anonymous et al., 2007f] analyzer which can measure blood chemistry (glucose, blood

gases, electrolytes, urea and more) in whole blood. The iSTAT uses a combination of

capillary action and external pressure for fluidic actuation.

Apart from this, in the microchannel based heterogeneous immunoassays the

molecular recognition elements (antibodies) are usually immobilized onto the channel

walls or on to micro beads. Surface immobilization further requires additional

microfabrication processing steps and suffers from poor reproducibility and reliability.

Binding the antibodies onto the microbeads in contrast, obviates the need for surface

modification of the channels while also offering significantly larger surface area for

binding and a higher sensitivity [Verpoorte et al, 2003]. Being mobile, microbeads can also

be circulated in the liquid by external means, to accelerate mass transport and further

reduce the reaction times. Bead based systems still require some mechanism to hold them

in place during the separation or washing step in the assays.

The most popular approach to immobilize the beads has been to use microfabricated

physical barriers that retain the beads while allowing the solution to pass through. Sato et

al. [2000] used a microfabricated dam structure to localize beads in an immunoassay

system to detect IgG, carcinoembryonic antigen [Sato et al, 2001], and interferon-gamma

[Sato et al, 2001]. Assay times were reduced by a factor of 70 from 24 hours to 20 minutes.

Moorthy et al. [2004] described the design and fabrication of a microfluidic system for

assaying botulinium toxoid by sandwich ELISA directly from whole blood. The device

incorporated a porous filter to separate the serum from the blood cells, and avidin-agarose

beads held by a filter membrane as the solid phase. Christodoulides et al. [2002] fabricated

micro machined pits to entrap beads in an immunoassay chip developed for cardiac

markers. An alternative approach for separation is to use paramagnetic beads localized by

an external magnetic field. This configuration permits highly flexible microfluidic

architecture, since the fluidic portion is isolated from the separation mechanism. Hayes et

al. [2001] developed a flow based small volume (<1 µL) micro immunoassay and

14

Figure 2.4 Gyros compact disc microfluidic chip

Figure 2.5 Micronics microfluidic card

15

demonstrated direct interaction of FITC and anti-FITC coated magnetic beads where 90%

of the maximum signal was reached in 3 minutes.

Heterogeneous sandwich assays for parathyroid hormone and interleukin-5 were

also demonstrated [Hayes et al, 2001]. Farrell et al. [2004] developed a magnetic bead

based immunoassay for mouse IgG using electrochemical detection which was eventually

integrated with a generic microfluidic platform by Ahn et al. [2002]. The total time

required for the assay was 20 minutes using 10 µL of sample per assay. An electromagnet

was used to immobilize the magnetic beads and the detection strategy employed to detect

the signal was electrochemical detection. The whole device was based on continuous

microfluidics which limited its use to specific protocols. Furthermore, robust tubing was

involved to pump all the reagents through valves into the microfluidic device.

Despite the commercial existence of a few lab-on-a-chip devices to perform

immunoassays, none of the aforementioned commercial devices perform bead based

immunoassays. This is because of the difficulty in mobilizing the magnetic beads [Gijs et

al, 2004]. Magnetic separation is different from magnetic transport since the magnetic

beads are separated using magnetic force, but are transported using liquid flow. In magnetic

transport, magnetic forces effectively transport the particles; in general the magnetic forces

and fields required for transport are much greater than those required for separation.

2.6 Diagnosis of Insulin and IL-6

2.6.1 Interleukin-6

Interleukin-6 (IL-6) is a multifunctional protein that regulates the immune response, acute

phase reactions, and hematopoiesis. IL-6 is produced by lymphoid and non- lymphoid cells,

and by normal and transformed cells, including T-cells, B cells, monocytes, fibroblasts,

vascular endothelial cells, cardiac myxomas, bladder cell carcinomas, myelomas,

astrocytomas and glioblastomas. The production of IL-6 in these various cells is regulated

either positively or negatively by a variety of signals including mitogens, antigenic

stimulation, lipo polysachcharides, IL-1 and viruses. On the basis of its various activities,

IL-6 has also been called interferon- (IFN- 2), 26kDa protein, B-cell stimulatory factor-2

(BSF-2), hybridoma/plasmacytoma growth factor, hepatocyte stimulating factor, and

macrophage-granulocyte inducing factor 2A (MGI-2A). A lot of reviews have been

16

published on interleukin-6 [Kishimoto, 1992; Hirano, 1990; Hirano, 1992; Hirano, 1990].

Disruption of IL-6 regulation might affect the immune response and consequently induce

the immune-mediated inflammatory diseases such as rheumatoid arthritis, systemic

juvenile idiopathic arthritis, Castleman disease and Crohn’s disease. Over production of IL-

6 [Nishimoto, 2006] also contributes to the development of malignant diseases such as

multiple myeloma and renal cancer. These proteins exist in really low concentrations in the

blood of the diseased and require very sensitive assays to detect such low concentrations.

2.6.2 Insulin

Insulin is a hormone and like many hormones, it is a protein. It is secreted by

groups of cells within the pancreas called islet cells [Wright et al., 1968]. Carbohydrates (or

sugars) are adsorbed from the intestines into the blood stream after a meal. Insulin is then

secreted by the pancreas in response to this detected increase in blood sugar. Most cells of

the body have insulin receptors which bind the insulin which is in the circulation. When a

cell has insulin attached to its surface, the cell activates other receptors designed to absorb

glucose from the blood stream into the inside of the cell.

Without insulin one can eat lots of food and still remain in a state of starvation since

many of the body cells cannot access the calories contained in the glucose very well

without the action of insulin. Those who develop a deficiency of insulin are considered

patients with Type 1 Diabetes where they should have it replaced by insulin shots or

pumps. More commonly people develop insulin resistance (Type 2 Diabates) rather than a

true deficiency of insulin. In this case, the insulin levels in the blood are similar or even

little higher than normal, non-diabetic individuals. However many cells of Type 2 diabetics

respond sluggishly to the insulin they make and therefore their cells cannot absorb sugar

molecules well. This leads to blood sugar levels which run higher than normal.

2.6.3 Current diagnostic techniques to detect Insulin and IL-6

Although there exist different kinds of immunoassays to determine the

concentration of IL-6, there exists only one commercial device which can perform an

immunoassay on-chip [Nobuo et al., 2005]. However, the device is based on continuous

17

microfluidic format relying on centrifugal forces to transport liquid. The same is the case

with Insulin where only one or two devices exist in the commercial arena. However since

the concentration of these analytes in the human physiological fluids is very low, it requires

the sensitive assays involving usage of magnetic beads to perform the assays. The existing

formats of microfluidic devices cannot perform assays involving magnetically responsive

beads because of difficulty in transporting the beads on a microscale. In fact there is no

commercial device which performs immunoassays involving magnetically responsive

beads. Hence a digital microfluidic lab-on-a-chip which can transport discrete droplets of

liquid by manipulating the surface tension is designed and developed in this thesis to

perform magnetic immunoassays on Insulin and IL-6.

2.7 Chapter Summary

In this chapter, the applications of microfluidics in clinical diagnostics are described.

Different kinds of automated immunoassay analyzers were explained and the disadvantages

of these devices in a point-of-care setting were explained. Different kinds of existing

commercial microfluidic devices based on continuous microfluidics are explained. The

reasons for the need for a new microfluidic device were explained. Existing detection

strategies for Insulin and IL-6 were explained.

18

CHAPTER 3

DIGITAL MICROFLUIDICS

3.1 Droplet based microfludics

In the previous chapter, the challenges in performing a magnetic immunoassay in a

continuous microfluidic platform were described. An alternative to continuous-flow

microfluidic systems was developed at Duke University using discrete droplets of liquid

which can be transported without any fixed microchannels based on surface tension

gradients. This approach has several advantages over the continuous-flow systems the most

important being reconfigurability and scalability of architecture [Pollack et al., 2002].

Established assays and chemistry protocol can be easily scaled down by using discrete

droplet based protocols since they are functionally equivalent to bench-scale wet chemistry.

Electrowetting and dielectrophoresis are the two most common actuation strategies to

manipulate droplets of liquid. Though both of these phenomena have electrostatic origin

electrowetting is fundamentally a contact line effect while dielectrophoresis is a body

effect. Furthermore, in the case of dielectrophoresis, which uses high frequency AC voltage

(>50 Hz), significant Joule heating occurs in samples even with moderate ionic strengths.

Hence, electrowetting appears to be applicable to a wider range of solutions and reagents

when compared to dielectrophoresis.

3.2 Surface tension driven flow

It has long been known that surface tension gradients can produce bulk flows of liquid

films or droplets. The oldest and most often cited example of this is the “tears of wine”

effect where wine droplets are driven along the surface of a wine glass due to surface

tension gradients arising from the evaporation of alcohol. Such flows arising from

chemically or thermally induced surface tension gradients within a liquid meniscus are

known as Marangoni flows and are important in many industrial applications [Probstein,

1989]. We first consider the forces acting on a liquid droplet resting on a non-wetting

19

surface in the absence of gravity. The forces at the solid-liquid-vapor interface are

described by Young’s equation in which the surface energies if each of the three interfaces

are represented by surface tensions acting along their associated surfaces (Figure 3.1). At

equilibrium the forces are balanced and the contact angle θ between the droplet and the

solid is determined by balancing the three interfacial tensions,

SLSVLV γγθγ −=cos (3.1)

where LV, SV and SL are the liquid-vapor, solid-vapor and solid-liquid surface energies

respectively [Adamson et al., 1997]. When a droplet is in contact with a surface having a

surface energy as shown in Figure 3.2, a net force arises between the ends of the droplet

due to the out-of-balance surface tension forces acting along the contact line. This net force

may induce bulk flow of the droplet. One of the earliest and most straightforward

demonstrations of this effect was demonstrated by Chaudhary et al [1992]. They exposed a

silicon wafer to a diffusing front of decyltrichlorosilane vapor to produce a smooth gradient

of surface energy. When droplets of water were placed at the more hydrophobic end of the

wafer inclined at 15º to the horizontal, droplets moved uphill towards the more hydrophilic

end at a rate of 1-2 mm/s. However the presence of surface tension alone is not sufficient to

induce motion of the droplets if significant hysterisis in contact angles is present. Contact

angle hystersis is essentially the observation that real contact lines are characterized not by

a single equilibrium contact angle but by a range of values [Adamson et al., 1997; Dussan,

1979]. A recently advanced contact line exhibits higher contact angle θA than a recently

receded one with contact angle θR. At equilibrium the contact angle may assume any value

between these limits without disturbing the position of the contact line. Thus, some

additional force is required to initiate and sustain the motion of the droplet. An everyday

example is a rain drop attached to the surface of a vertical window pane where the

difference in angles between the ends of the droplet creates a force that opposes the

downward motion due to gravity. At present, the origin of contact angle hysterisis is poorly

understood, but is clearly associated with the surface roughness and heterogeneity of the

surface.

20

θ

γSLγSV

γLV

LiquidFiller medium (oil

or vapor)

Solid

Figure 3.1 Cross section of equilibrium forces acting on one side of non-wetting droplet in contact with a horizontal surface

Figure 3.2 Cross section of a droplet resting on a surface having a gradient of surface energy. Droplet shapes are exaggerated to show equilibrium contact angles on each side.

The arrows indicate the direction of motion

21

3.3 Electrowetting based droplet actuation

The term electrowetting was first introduced to describe an effect proposed for designing a

new display device [Beni et al., 1981]. Electrowetting was initially defined as “the change

in the solid electrolyte contact angle due to an applied potential difference between the

solid and the electrolyte” [Beni et al., 1981]. There have been significant developments in

this technology since then and a large number of microdevices have been devised actuate

discrete droplets of liquid. This actuation of droplets in a micro device was termed as

“Digital microfluidics” where the liquid is transported by modulating the interfacial tension

between the polar liquid phase and a solid electrode by the application of an electric

potential between the two. The solid electrode is insulated in order to prevent the liquid

from electrolyzing. In a typical electrowetting setup, the droplet is sandwiched between

two surfaces, one of which is a patterned electrode and the other a continuous ground

plane, which could be an Indium tin oxide (ITO) coated glass plate and is completely

surrounded by a filler fluid. There exists a gap between the two surfaces sandwiching the

droplet which also defines the height of the droplet. The droplet is further insulated from

the electrode by a dielectric and all the surfaces are hydrophobized. Figure 3.1 shows the

wetting of an open droplet on an electrode with an insulation layer.

3.3.1 Theory of electrowetting

Consider a conducting droplet which is resting on a non-wetting solid insulator of thickness

d and relative dielectric constant of εr as shown in Figure 3.3. If a voltage V is applied

between a wire electrode inside the droplet and an electrode underneath the insulator the

charge will be stored in the effective capacitor formed between the droplet and the

electrode underneath. The electric energy stored in the capacitor is directly proportional to

the area A of the base of the droplet assuming that the interface is uncharged in the absence

of any potential [Adamson et al., 1997]. By the parallel plate approximation,

20

2V

d

AE rεε

= (3.2)

This electrostatic energy directly modifies the solid-liquid interfacial tension SL so that

22

20

2)0()( V

dV r

SLLS

εεγγ −= (3.3)

which is essentially Lipmann’s equation originally derived for a mercury-electrolyte

ineterface with charge separation across a diffuse region in the liquid rather than a solid

insulator. The contact angle of the liquid with the solid is modified according to equation

3.1 giving

20

2)0(cos)(cos V

dV

LV

r

γεεθθ += (3.4)

Hence, the voltage reduces the contact angle and the droplet spreads or wets the surface. It

can also be explained that the surface becomes more hydrophilic on application of voltage

to the droplet and this effect is theoretically independent of polarity and frequency of the

effective voltage. The electrode underneath the droplet can be divided into an array of

multiple, independently controlled electrodes to provide the required asymmetry. The

droplet volume is slightly larger than the pitch of the electrodes to ensure overlap of the

droplet with the adjacent electrodes as shown in Figure 3.4. This enables droplet motion

when a potential difference is applied between the electrode underneath the droplet and the

adjacent electrodes, by establishing a surface energy gradient which drives the droplet onto

the charged electrode. The existence of independently controllable electrodes wherein the

force is directly applied at the meniscus of the liquid makes it a system with no fixed

channels and structures to contain the liquid. Hence, reconfigurability within the liquid

layer is facilitated since the reservoirs, channels and mixing structures exist in a virtual

sense.

Figure 3.4 shows a side and top view of a droplet in a basic electrowetting setup.

The metal electrodes are defined in chrome or indium tin oxide (ITO). The insulator is a

parylene film whose thickness can be varied between 5 and 12 µm depending on the

voltage applied on the electrode solutions, which in turn depends on the solutions that are

transported on the chips. The insulator is hydrophobized by coating a film of the

23

Droplet

V

Insulation

Electrode

θ

Figure 3.3 The electrowetting effect

Silicone oil Droplet

Insulation layer

Electrodes 1 2 3 4

Hydrophobic layer

Top plate

Droplet

Top plate

Side view Top view

1 2 3 4

Figure 3.4 Top and side views of a basic electrowetting setup

24

fluoropolymer Teflon AF. A low viscosity (1.5 cSt) silicone oil is used as the filler

medium. Each patterned electrode on the array is considered to be a discrete position for

the droplet to occupy when the electrode is activated. All the fundamental fluidic

operations like transport, merging, mixing, and splitting of droplets can be performed by a

single instruction of activating the electrode. Due to the similarity to the functioning of a

digital microprocessor, this approach is referred to as “Digital microfluidics”.

3.4 Electrowetting based lab-on-a-chip for point of care testing

The basic fluidic operations involved in any application on a digital microfluidic lab-on-a-

chip can be broadly categorized as sample preparation, droplet dispensing from a larger

volume of liquid [Pollack et al., 2001; Ren et al., 2004], droplet splitting, droplet

manipulation [Pollack, 2002], droplet mixers [Paik et al., 2003a; Paik et al, 2003b; Paik et

al, 2003c]. All these basic droplet operations have been well characterized by previous

researchers. Furthermore applications in clinical chemistry [Srinivasan et al, 2005] such as

glucose assay on lab-on-a-chip and on chip cooling [Paik et al, 2006] based upon the

digital microfluidic architecture has also been studied. Each of the above is described

briefly in the following paragraphs.

3.4.1 Droplet dispensing

Dispensing of unit sized 700nL droplets of 0.1M potassium chloride (KCl) solution from a

larger (reservoir) initial droplet was demonstrated by Pollack et al [2001]. The droplets

were formed as shown in Figure 3.5 by extending a finger of liquid and switching off the

intermediate electrodes to ensure the finger breaks off creating a unit sized droplet slightly

bigger than the electrode. Droplets were also dispensed from an external source using a

pipette to inject liquid onto an energized electrode and withdrawing the pipette to form a

droplet on the energized electrode [Pollack et al, 2001]. Dispensing of micro liter droplets

using external pressure from a reservoir off chip was demonstrated using capacitance

metering by Hong et al [2004]. The volume reproducibility was very good and the volume

variation was reported to be less than 2%.

25

3.4.2 Droplet transport

Transport of nanoliter droplets of 0.1M KCl at a maximum average speed of 10 cm/sec

using voltages less than 60V (Figure 3.6) was shown by Pollack et al. (2002). Droplets

were manipulated in an open system in air and in immiscible filler fluids such as silicone

oil. The motion of the droplet was independent of molar concentration and viscosity of the

liquid being transported. However the motion of the droplets was significantly inhibited

with increasing viscosity of the filler fluid. All the experiments described above utilized

non biological samples and used manually dispensed droplets. Srinivasan (2005)

demonstrated the bio-compatibility of the digital microfluidic lab-on-a-chip by transporting

various proteins and enzymes.

3.4.3 Sample preparation and droplet dilution

An on-chip serial dilution scheme was demonstrated by Ren et al. (2003). The dilution

scheme used was a serial dilution scheme with mixing and splitting of two droplets of

different concentrations to obtain two droplets with an intermediate concentration. A range

of dilution factors were tried and it was reported that the errors increased exponentially

with each dilution step and the total error after the Nth dilution can be modeled by the

equation

EN= 1-(1-E0)N

(3.5)

where EN was the error after N dilution steps (dilution of 2N) and E0 is the error in the basic

two dilution step.

3.4.4 Efficiency and speed of droplet mixing

Droplet mixing on any microfluidic system is a challenge because of highly laminar flow

conditions and flow reversibility. Droplet mixing was extensively studied by Paik et al..

(2003a). It was reported that purely diffusive mixing took about 90 seconds whereas the

time of mixing was reduced to 4.6 seconds by linearly oscillating the droplets over 4

electrodes. Mixing times were further reduced to less than 3 seconds using a 2-dimensional

26

Figure 3.5 Dispensing of KCl droplets from a reservoir

Figure 3.6 Transport of a droplet on a 2 phase bus

27

array mixer, which reduced the effects of flow reversibility. Further reduction in the mixing

time was obtained (less than 2 seconds) by employing a split and merge sequence of

droplets.

3.4.5 Biocompatibility of electrowetting platform

All the droplet operations described above used simple aqueous salt solution or an

electrolyte such as 0.1M KCl. However mobility of proteins and other human physiological

fluids becomes increasingly difficult to handle because of the tendency of proteins to

adsorb on to any surface they come in contact with. It was also reported that the adsorption

is more pronounced in the case of hydrophobic surfaces. In most cases the effect is

irreversible under normal conditions [Yoon et al., 2003]. Usually the hydrophobic coating

that is used is Teflon AF coating. Hence any contact between the Teflon AF surface and the

protein droplet would contaminate the surface, which is detrimental to transport because

electrowetting works on the principle of modifying the wettability of the hydrophobic

surface. Although static electrowetting (static change in electrowetting by application of

potential) of 4 µg/mL of bovine serum albumin (BSA) was demonstrated [Yoon et al.,

2003] in a very limited sense, dynamic transport has not been shown. Silicone oil, which

has low surface tension and enough spreading properties, was chosen as the filler medium

to minimize the adsorption of the proteins on the surface. Although a layer of oil exists in -

between the droplet and the electrode, the stability of the film is yet to be extensively

characterized because the stability of the film decreases with decreasing interfacial tension

between the droplet and the oil. The less stable the oil layer, the more the adsorption of the

proteins onto the surface. Furthermore, as the interfacial tension decreases, the proteins wet

the hydrophobic surfaces more than normal, squeezing the oil layer out due to electrostatic

pressures [Srinivasan et al., 2005]. Hence the actuation voltage plays an important role in

the protein mobility. Similar difficulty in transportability due to adsorption was observed in

the case of solutions with different concentrations of surfactants. It can be seen from the

Figures 3.9 and 3.10 (Schonherr et al., 1971) that the surface tension and the contact angle

reduce with increasing concentration of the surfactant. This enhances the wetting of the

droplet on the surface thereby increasing the adsorption onto the surface, which in turn

makes the transport more difficult.

28

Srinivasan et al. [2004] demonstrated the biocompatibility of the digital microfluidic

platform by transporting human whole blood, plasma, serum, urine, saliva, sweat and tears.

Figure 3.7 shows a plot of actuation voltage utilized and the maximum switching frequency

at that particular voltage. Due to the discrete nature of the digital microfluidic system, the

maximum switching frequency which is defined as the highest rate at which a droplet can

be moved across two adjacent electrodes is a measure of the transport performance of the

system. The average speed of the droplet is defined as the product of switching frequency

and the electrode pitch. Therefore, the switching frequency is inversely proportional to the

electrode pitch and can be increased by physically scaling down the system. It can be seen

that different voltages are required depending on the interfacial tension of the liquid droplet

and the oil film. Figure 3.8 shows a droplet of blood being transported on a transparent ITO

electrode.

Droplet formation of proteins [Srinivasan et al., 2004] has also been demonstrated

and it was reported that reliable on chip dispensing was possible up to concentrations of

0.01 mg/mL for BSA. However for solutions with higher concentrations of BSA (>=0.1

mg/mL), it was not possible to dispense the droplets on chip despite the fact that individual

droplets containing 10 mg/mL of BSA are transportable which might be because of higher

protein adsorption in the reservoir, which is larger in area. Physiological fluids such as

serum, plasma and other enzymatic reagents were dispensed reliably [Srinivasan et al.,

2005]. However, droplet dispensing did not work for whole blood. Apart from establishing

the transportability of various physiological fluids on the digital microfludic lab-on-a-chip

Srinivasan et al. [2005] evaluated the applicability of the platform to clinical chemistry by

performing glucose assays using standard solutions and compared them with the results

obtained using a reference method on a spectrophotometer. It was reported that there is no

loss of enzyme activity on the electrowetting platform. Protein stamping for MALDI

spectrometry was also performed on this platform using electrowetting [Srinivasan et al.,

2004]. Paik et al. (2006) developed an adaptive cooling architecture based on the digital

microfluidic platform to demonstrate the feasibility to adaptively perform thermal

management to dynamically cool hot-spots for applications such as Polymerase chain

29

Figure 3.7 Frequency voltage curves for various physiological fluids

Figure 3.8 Transport of blood droplet on transparent ITO electrodes

30

reaction (PCR) on chip. All the applications described above were done with human

physiological samples and different concentrations of proteins.

The overall objective of this thesis is to design and build a nano-liter lab-on-a-chip

to perform an integrated and automated immunoassay involving magnetically responsive

beads. This would involve mobility of magnetic beads in a nano scale, which has not been

studied before.

3.5 Chapter Summary

Digital microfluidics for manipulating discrete droplets of liquid was presented in this

chapter. The concept of electrowetting to manipulate the liquid droplets by application of

voltage was explained. A brief literature survey of the droplet operations that can be

performed on the digital microfluidic platform and the biocompatility of various analytes

was presented.

31

CHAPTER 4 LAB-ON-A-CHIP DESIGN, FABRICATION AND

TESTING

In this chapter, a generic architecture is developed to perform an immunoassay on a lab-on-

a-chip using a droplet-based approach. This lab-on-a chip is used to establish the proof-of-

concept of performing an immunoassay involving magnetically responsive beads using

droplet based transport. Fabrication procedures for this lab-on-a-chip are discussed and the

fabricated chip with the magnetic beads and the reagents used in the immunoassay are

tested. Furthermore, the detection instrumentation used for measurement of

chemiluminescence is also described in this chapter.

4.1 Lab-on-a-chip specifications

The basic requirements to perform an immunoassay on a droplet based lab-on-a-chip are

described in this section. The functional components of the lab-on-a-chip are the droplet

generation units and droplet pathways for transport, mixing, incubation and washing of the

magnetic beads, which is the most important step to perform a magnetic immunoassay.

Figure 4.1 shows the design of a droplet based electrowetting lab-on-a-chip developed as a

part of this dissertation to perform an automated immunoassay using magnetically

responsive beads.

4.2 Fabrication of Lab-on-a-chip and system assembly

4.2.1 Chip fabrication (Figure 4.2)

The electrowetting chips were fabricated using a commercially available photomask

manufacturing process. Electrode patterns were photolithographically imaged and etched

on a chrome-coated photomask blank. The base material was 0.060" thick soda lime glass

and the thickness of the chrome layer was 840A. Photosciences Inc., Torrance, CA was

used as the vendor to fabricate the chips.

32

Figure 4.1 Lab-on-a-chip design to perform immunoassay involving magnetic beads

Reservoirs

Electrodes

Connector pads

33

Parylene C was used as the dielectric on all the chips. The Parylene C coating process is a

conformal vapor deposition process which produces a high-quality pinhole free dielectric

layer. The Parylene C was coated at the SMIF Facility at Duke University, Durham, NC or

Paratronix Inc, MA.

Teflon AF was used as the hydrophobic coating on all the chips. Teflon AF is an

amorphous fluoropolymer, which is soluble in perfluorinated solvents making it suitable

for application as thin films by either dip coating or spin coating. A 200nm-1000nm thick

Teflon AF layer was typically used as the hydrophobic coating on the electrowetting chips.

Indium-tin-oxide (ITO) coated glass plates were used as the cover plates to create the

droplet sandwich. The ITO plates were also coated with a thin layer of Teflon AF.

4.2.2 Top plate fabrication

An indium-tin oxide (ITO) or poly-carbonate plastic acted as the ground plane. The

indium-tin oxide slides were obtained from Delta Technologies. Polycarbonate plastic

pieces were coated with ITO by Genvac Aerospace. Holes were drilled in the ITO coated

top plate such that they aligned with the reservoirs of the chip. The top plates were later

cleaned by sonicating in Isopropanol and then air dried. The ITO plates were

hydrophobized by spin coating or dip coating (<100nm thick film) with 1% Teflon AF. The

glass plates were then baked at 180º C for 45 minutes to remove the solvent.

4.2.3 System assembly

The droplet based lab-on-a-chip and the top plate were assembled such that the holes on the

top plate were aligned perfectly with the reservoirs at the end of the electrode lines. The

chip and the top plate were clamped and held down by microscope clips on the sides of the

top plate. A gap height of 300 µm was used which was created by placing a 300 µm shim in

between the top plate and the chip. Electrical connections to the contacts were made using

a 22-pin SOIC test clip (Pomona electronics). Voltages were applied to the test clip using

an electronic controller which was essentially an array of high-voltage switches. The state

of the switches (ON or OFF) was controlled by custom written software through the

parallel port or USB port of the computer.

34

Glass substrate

Photoresist

Photo

mask

Photoresist

Glass substrate

Glass substrate

Glass substrate

Chrome deposition

Glass substrate

Stripping the photoresist

Parylene coating

Teflon coating

Glass substrateGlass substrate

Glass substrateGlass substrate

Glass substrate

1

2

3

4

5

6

7

Figure 4.2 Fabrication process for droplet based lab-on-a-chip for immunoassays

35

4.3 Detection Instrumentation

Detection in electrowetting can be done using optical techniques because of the transparent

nature of the materials involved in the chip. Srinivasan et al., 2005 developed an optical

absorbance measurement system consisting of an LED and a photodiode which is

integrated with the electrowetting device to monitor the color obtained during the glucose

assays. However, for the immunoassays, since the concentrations to be detected are very

low, such low levels of color cannot be detected at path lengths equivalent to the gap height

used (300µm). Although measurement of fluorescence is a good method of detection,

setting up of such a detection technique integrated to the electrowetting device is difficult.

The other detection technique that is most commonly used for measuring the

concentrations of analytes in immunoassays is chemiluminescence. Chemiluminescence is

the emission of light without emission of heat as a result of a chemical reaction.

[A] +[B] [ ] [AB] + LIGHT

The decay of the excited state to a lower energy level is responsible for the emission of

light. Light detection technology is a powerful tool that provides deeper understanding of

more sophisticated phenomena occurring in the reaction. Measurement of light offers

unique advantages: non destructive analysis of a substance, high speed performance and

extremely high detectability. Recently fields as scientific measurement, medical diagnosis

and treatment, high energy physics, spectroscopy and biotechnology require the

development of photodetectors that require extremely high performance. Photodetectors or

light sensors can be broadly classified into three major categories based on their operating

principle: external photoelectric effect, internal photoelectric effect and thermal types. The

external photoelectric effect is a phenomena in which when light strikes a metal or

semiconductor placed in a vacuum, electrons are emitted from its surface into the vacuum.

Photomultiplier tubes (often abbreviated as PMT) make use of this external photoelectric

effect (shown in Figure 4.3) and are superior in response and sensitivity (low-light level

detection). Hamamatsu’s PMT is used for this research to detect the chemiluminescence

emitted during the immunoassays. The detection assembly is shown in Figure 4.4 and the

connection circuit to detect the chemiluminescence is shown in Figure 4.5

Intermediate complex

36

Figure 4.3 Working of a PMT based on external photoelectric effect

37

Teflon AF layer

Parylene layer

Chrome electrodes

Glass

Glass

ITO layer

Detector

board

Photo multiplier tube

Droplet1.5cSt Silicone oilH

Monitor

CCD

camera

Computer

Figure 4.4 Detection instrumentation for chemiluminescent detection in the droplet based lab-on-a-chip

38

DAC

PMT

R

i

Signal out

LTC 1152

ADC

Microcontroller Computer

1.2 V

ref

Gain

voltage

Input light

Figure 4.5 Circuit for chemiluminescent detection of the signal using a PMT

39

4.4 Design of the reservoir and the droplet pathways

Srinivasan et al.[2005] described the design of the electrode for the reservoir for

performing a glucose assay on a lab-on-a-chip. Figure 4.6 depicts the various stages in

dispensing a droplet from the reservoir. Dispensing occurs in the following three steps:

[Ren et al., 2003]

1. A liquid column is extruded from the reservoir by activating a series of electrodes

adjacent to it as shown in Figure 4.6 (1, 2, 3, and 4).

2. Once the column overlaps the electrode where the droplet is to be formed, all the

remaining electrodes are deactivated to form a neck in the column.

3. The electrode in the reservoir is then activated to pull back the liquid causing the

neck to break completely and form a droplet almost equal to the size of the

electrode.

The reservoir electrode in this design has a tapering-pull back electrode (wider at the

dispensing end) to ensure that the liquid always stayed at the dispensing end of the

reservoir. The different parameters that would affect the reproducibility and reliability of

the dispensing process are the reservoir shape and size, shape and size of the pull back

electrode, size of the unit electrode which would decide the size of the unit droplet and the

spacer thickness which would decide the volume of each droplet. The choice and effect of

each of these parameters is discussed in the following paragraphs [Ren et al., 2004].

Electrode size- The electrode size is chosen to be 1250 µm because of the

sensitivity required in immunoassays. However the electrode pitch can be reduced once the

proof of concept of an immunoassay on lab-on-a-chip has been established to further

reduce the sample size.

Spacer thickness- Previous results [Ren et al., 2003; Cho et al., 2003] indicate that

the droplet dispensing for a water-silicone oil system requires a droplet aspect ratio

(diameter: height) greater than or equal to 5. However to transport droplets containing

magnetic beads and to immobilize magnetic beads within a droplet the gap height should

be >250 µm for efficient attraction of the beads. Hence a gap height of 300 µm is chosen.

40

Figure 4.6 Scheme of dispensing a droplet from a reservoir on a lab-on-a-chip made with Printed Circuit Board (PCB) technology

1 2 3

654

ON

ON OFF

OFF ON

ON

ON ON

ON ON

ON

ON ON

ON ON ON ON

ON ON ON

OFF

OFF

OFF

OFF OFF

OFF OFF

OFF

ON

ON

1

41

For this electrode pitch and gap height the volume of each unit droplet would be

approximately 500 nL. This would also provide enough sensitivity for very low

concentration of analytes. Apart from the design parameters discussed above, other

parameters which would affect dispensing and transport of the liquid include voltage

applied and the volume of the liquid pipetted onto the lab-on-a-chip. Furthermore the

number of electrodes in a single array is another important parameter to perform a

magnetic immunoassay on chip to perform efficient splitting of the supernatant after

immobilizing the magnetic beads.

4.5 Biocompatibility of the lab-on-a-chips

The fabricated lab-on-a-chips were tested with a wide range of reagents that will be used to

perform the magnetic immunoassay. An extensive list of the reagents tested on the lab-on-

a-chip is given below:

1. Antibodies for Insulin and Interleukin-6 (IL-6)

2. Different concentrations of Insulin and IL-6

3. Magnetically responsive beads

4. Different concentrations of Tween® 20

5. Different concentrations of Triton X 15

6. Different concentrations of Bovine serum albumin (BSA)

7. Serum

8. Different concentrations of Horse radish peroxidase (HRP) enzyme

9. Different concentrations of Alkaline phosphatase (ALP) enzyme

10. Lumigen APS-5 (chemiluminescent substrate for ALP enzyme)

11. Lumigen Ultra PS-Atto (Chemiluminescent substrate for HRP enzyme)

Experimental protocol for testing transportability and resuspension of magnetic beads-

1µL of the reagent is pipetted manually on to one of the electrodes on the lab-on-a-chip and

is sandwiched using an ITO top plate for grounding. The gap height used was 300 µm

which was filled with 1.5 cSt Silicone oil. The droplet was transported across 5 electrodes

to and fro at different frequencies and different voltages until the droplet stops to transport

completely.

42

Observations- The reagents mentioned above are proteins and surfactants and have a lower

surface tension when compared to water. Hence they have the tendency to adsorb to the

surface of the chip and render the chip to be permanently hydrophilic. It can be deduced

from Eq 3.4 that as the interfacial tension between the solid and the liquid droplet

decreases, the contact angle reduces and the droplet wets or spreads more upon application

of voltage. This effect will be enhanced upon application of voltage. It was observed that

different voltages were required for different solutions based on the surface tension. The oil

film and the interfacial tension of the droplet-oil interface are yet to be characterized

extensively for different reagents. The maximum concentration of BSA, HRP that were

transportable was 10 mg/mL and 2 mg/mL respectively. The magnetically responsive beads

transported really well on the droplet based lab-on-a-chip. Figure 4.7 shows the transport of

the magnetically responsive beads being transported across five electrodes with a magnet

underneath the electrodes to attract the magnets. In stage 1 of Figure 4.7, the beads were

aggregated because of the effect of the magnetic field; however they were completely

resuspended in the buffer after shuttling across 5 electrodes. The maximum concentration

of Tween ® 20 in Phosphate buffered saline (PBS) that was transportable without major

difficulty was 0.01%. The physiological fluids such as serum, plasma and blood also

transported very well. However, since all the above tested solutions have proteins and

surfactants at different levels, the interfacial tension between the droplet and the chip

surface reduces which enhanced adsorption onto the surface of the chip and hinder the

movement of the proteins. The adsorption occurs mainly because of an unstable oil film

between the droplet and the chip surface.

Droplet dispensing- All the above mentioned reagents involved in the magnetic

immunoassay were dispensed. Physiological liquids such as serum (Figure 4.8) and plasma

were also dispensed. However dispensing became increasingly difficult with increasing

concentration of the protein. This is because of increased surface area on the reservoir

which enhanced surface adsorption thereby hindering transportability of the droplets by

making the surface permanently hydrophilic. Whole blood could not be dispensed from the

reservoir despite efficient transportability. This is because of different kind of proteins and

43

1 2 3

654

Figure 4.7 Transport of a unit droplet containing magnetically responsive beads across five electrodes

1 Tesla Magnet

44

ON

ON ON

OFF

ON

ONON

OFF

ON

ON

OFF

Figure 4.8 Dispensing and Transport of serum on Lab-on-a-chip

45

surfactants present in the blood which lowers the surface tension significantly enhancing

adsorption.

Droplet splitting- Splitting of larger slug of droplet into two equally sized droplets was

termed as symmetric splitting and splitting of a larger slug asymmetrically was termed as

asymmetric splitting. Washing of the magnetic beads would involve splitting of the

supernatant from the retained magnetic beads. Hence droplet splitting is the most important

operation to perform the magnetic immunoassay on chip which needs to be performed

consistently and redundantly. Asymmetric and symmetric splitting was demonstrated using

Tween® 20 droplets. It was observed that there was consistent and reproducible splitting at

the same point.

In all the aforementioned droplet operations involving proteins and surfactants of

different concentrations, the transportability was difficult after some time because of the

tendency of the proteins to adsorb to hydrophobic surfaces. It can also be seen from the

Figures 3.9 and 3.10 that the increase in surfactant concentration reduces the surface

tension of the droplet wetting the surface more. This happens because of the reduction in

the contact angle which further reduces by the application of voltage. (Eq 3.4)

Inorder to characterize the transportability extensively on a digital microfluidic

platform based on electrowetting, the interfacial tensions existing between the droplets

intended to transport and the solid surface need to be determined. The interfacial tensions

can be determined precisely by using a tensiometer. However because of non availability of

resources, the transportability of the reagents involved in the magnetic immunoassay was

characterized empirically using the frequency versus voltage curves. Hence inorder to

stabilize the oil film and maintain the difference in the surface tensions ( SV and SL), 0.1%

Triton® X 15 was added to the 1.5 cSt silicone oil which was used as the filler medium.

This improved the transportability of the proteins significantly. The time to failure was

increased from 30 minutes to more than one hour. Hence the 1.5 cSt Silicone oil with 0.1%

Triton® X 15 was used for this research as the filler fluid.

46

1 2

43

Figure 4.9 Asymmetric splitting of 0.01% Tween® 20 in PBS

1 2

Figure 4.10 Symmetric splitting of water droplet

47

4.6 Protein fouling on chips

It was explained by Srinivasan et al. [2004] that the proteins contaminate the surface of the

chip and makes them permanently hydrophilic. Hence the following experiment was

performed to quantify the amount of contamination.

Materials- 1 mg/mL HRP, Lumigen Ultra PS Atto (chemiluminescent substrate for

HRP enzyme), 1.5 cSt silicone oil, Test chip 606 coated with 12 µm parylene and dip

coated with 6 % Teflon AF, ITO top plate dip coated with Teflon AF, chemiluminescent

detection setup Figure 4.4.

Methods- 1 µL of 1 mg/mL HRP was manually pipetted on an electrode on the

fabricated lab-on-a-chip and sandwiched using an ITO coated top plate which acts as the

ground plane. The gap height used was 300 µm which was filled with 1.5 cSt silicone oil.

The droplet was transported over three electrodes for 30 minutes and is discarded. To

quantify the amount of enzyme that was being deposited on the electrodes, 1 µL of

Lumigen Ultra PS atto and transported once over the same three electrodes over which the

HRP was transported and parked on one of the electrodes. The chemiluminescence was

measured for 500 seconds and compared to the background. It was observed that there was

increase in the chemiluminescent signal which confirms the proposition that proteins

adsorb to the hydrophobic surface. The assay performed was a really sensitive assay by

transporting a 1mg/mL of HRP over for 30 minutes. However the signal increase is not

significantly high. However the above experiment establishes that proteins adsorb to

surfaces which would vary the signal. Hence new electrode lines and new chips were used

for different assays.

4.7 Material defects

The thickness of the parylene coating and the concentration of Teflon AF used in

the fabrication of the droplet based lab-on-a-chip were 5 µm and 1% respectively. The

main reliability concern in the chips was the parylene insulator. Electrolysis in the chips

due to insulator was eventually seen in all the chips exclusively in the reservoir (Figure

48

Quantification of enzyme contamination on chip

330000

340000

350000

360000

370000

380000

390000

400000

410000

420000

0 100 200 300 400 500 600

Time (seconds)

Ch

em

ilu

min

esce

nce

(A

DC

co

un

ts)

Figure 4.11 Fouling of HRP enzyme on chip

Fouled

chip

49

4.12) and at the point of splitting (Figure 4.13). Though the exact reason for the breakdown

is unknown at this time it is hypothesized that this is due to the mechanical cracking or

failure of the parylene film at the gasket-electrode junction in the reservoir which exposes

the metal to liquid and causes electrolysis. This eventual breakdown of the insulator limited

the time duration of single automated experiments. The experiments lasted only for 15-20

minutes. Hence the next batch of chips were coated with 12 µm parylene and dip coated

with 6% Teflon AF. This solved the electrolysis problem and the chips lasted for more than

one hour without any electrolysis.

4.8 Chapter summary

The design and prototyping of a lab-on-a-chip for immunoassays was described in this

chapter. The biocompatibility of the droplet based electrowetting platform was established

by transporting different proteins applicable in an immunoassay including the mobilization

of magnetically responsive beads which does not exist in any of the commercially available

immunoassay analyzers. The insulator breakdown which caused catastrophic failures,

which ultimately determined the lifetime of the chip and duration of experiments was

improved by using a thicker parylene (12µm) and Teflon AF (6%) coatings. The protein

fouling on the lab-on-a-chip was quantified by transporting 1mg/mL of HRP for 30

minutes. Hence the fabricated lab-on-a-chip is capable of performing all the basic fluidic

operations such as droplet dispensing, transport, mixing and splitting.

50

Figure 4.12 Electrolysis in the reservoir

Figure 4.13 Electrolysis at the point of splitting

Electrolysis

Electrolysis

51

CHAPTER 5

ON-CHIP MAGNETIC IMMUNOASSAY

5.1 Magnetic Immunoassay

The immunoassay which utilizes magnetically responsive beads or spheres as the solid

phase is termed as “magnetic immunoassay”. Performing a magnetic immunoassay on chip

involves three basic steps: (i) affinity capture, (ii) separation and (iii) detection. The

separation step, which separates the magnetic beads with the antibody-antigen complex

from the excess/unreacted reagents, is the most difficult operation to perform on chip. The

washing step involves immobilizing the magnetically responsive beads at a single place

and removes most of the excess/unreacted supernatant. This process has to be repeated

until there is substantially no signal from the supernatant.

The washing step has to be performed in such a way as to

• Avoid permanent clumping or aggregation of the magnetic beads

• Capture and immobilize substantially all of the magnetically responsive beads

within a single droplet during a droplet splitting operation,

• Ensure immobilization and retention of substantially all of the magnetically

responsive beads,

• Ensure resuspension of all of the magnetic beads within the droplet with no

significant clumping or aggregation of the beads upon completion of washing

process.

There are several parameters that affect the efficiency of washing the magnetic beads

which would further influence the result of the immunoassay. Hence washing of the

magnetic beads is the most important step to perform the magnetic immunoassay on chip.

52

5.2 Super paramagnetic beads

Encapsulated super paramagnetic beads have found novel applications in the field of

biomedicine, drug delivery, cell separation and molecular biology [Technote#101, 1999;

Technot#301, 1999]. These super paramagnetic beads have the unique property of not

retaining any magnetism once the magnetic field is removed. This unique feature of these

magnetic beads makes them a perfect candidate for use as the solid phase in an

immunoassay. The advantages of the magnetic immunoassay are as follows:

• Separation is simple and fast

• Separation can be done with the basic laboratory equipment

• Can be used to determine fairly low antigen concentrations because of high surface

area available for binding on the magnetic beads

• Adaptability to automation

Super paramagnetic beads used for immunoassay are generally available in different sizes

ranging from a few hundreds of microns to nanometers. They are available labeled with

different proteins, enzymes and antibodies useful for different applications in the

immunoassay. The magnetic beads that were used in this research are uniform super

paramagnetic, polymer beads with streptavidin attached to the bead surface.

5.3 Parameters involved in washing of magnetic beads

The basic parameters involved in washing the magnetic beads on chip are listed below,

which are dependent mostly on material properties of the magnetic beads. Some of these

parameters were studied experimentally as described in the following paragraphs. Each of

these parameters was explored to establish a standard washing protocol to ensure very good

washing efficiency with minimal bead loss. High wash efficiency implies that there is very

little excess reagent left in the supernatant which would produce a secondary signal. Table

5.1 summarizes the parameters that were studied experimentally to produce an efficient

wash protocol.

53

Table 5.1 Parameters studied to characterize bead attraction

Parameter

1

2

3

4

5

Buffer PBS TBS PBS and TBS

with BSA

PBS with

0.005%

Tween 20

PBS with

0.01%

Tween 20

Magnetic pull

force (0.5 tesla

magnet)

1.25 lbs 5 lbs

Position of

magnet

Right

underneath

Over and

underneath

Quadrapole

arrangement

Concentration of

beads

Undiluted

stock

2 times

diluted

4 times diluted

5.3.1 Buffer system

The buffer system in which the magnetically responsive beads are suspended plays a very

important factor in the bead mobilization. Different types of buffers with varied

concentrations of salt and surfactant were used to suspend the magnetic beads and the

attraction and aggregation of beads were studied under magnetic field by taking periodic

images of the magnetic beads and analyzing the images using Image J software

(Anonymous et al., 2007 i). The following buffers were used to suspend the magnetic

beads:

• Phosphate buffered saline (PBS)

• Tris buffered saline (TBS)

• PBS and TBS with Bovine serum albumin (BSA)

• PBS and TBS with 0.005 % Tween® 20

• PBS and TBS with 0.01% Tween® 20

54

Materials- BioMAG streptavidin coated magnetic beads (cat# BM 551) obtained from

Bang’s Laboratories (Diameter-2.8µm) and Dynal® MyOneTM Streptavidin (Diameter-1.05

µm) (cat# 650.01) obtained from Dynal Biotech were used for this study. Tween® 20 was

obtained from Pierce. Stock solution of BSA (10 mg/ml) was provided by Glaxo

Smithkline (Durham, North Carolina) and diluted to 1 mg/ml, 0.1 mg/ml and 0.01 mg/ml in

PBS and TBS.

Experimental setup- Aliquots of the stock solution of the BioMAG streptavidin coated

magnetic beads were taken and diluted 4 times using the above mentioned buffers in

different tubes. This was done by pipetting the supernatant from the stock solution and

resuspending the beads with buffer. All the samples were further sonicated in an ultra-

sonicator to avoid any pre-clumping of the magnetic beads.1 µl of 1.5 cSt silicone oil was

pipetted on a Teflon coated glass slide and 1 µl of the beads were pipetted on the oil droplet

and a sandwich was created by placing a cover slip over the bead droplet. The gap height

used in the sandwich was 200 µm. An electrowetting setup was created for the droplet in

the sandwich between the Teflon coated glass slides. A 0.5 Tesla magnet was placed at a

distance of 5 mm from the bead droplet on the cover slip. The attraction of the beads was

observed under a microscope and periodic images were taken.

Observations- All the images (Figures 5.1, 5.2, 5.3) were processed by Image J software

(Anonymous et al., 2007 i) where the beads were considered as the dark pixels and the

supernatant as the bright pixels. A histogram of the whole area of the droplet was prepared

and the area of the dark pixels versus the bright pixels was compared. So the image with

more number of brighter pixels was considered to have better attraction of beads with all

the magnetic beads. It was seen that there was 40% better and quicker attraction by having

0.01% Tween 20 in the bead suspension. Details of the image analysis are given in

Appendix B.

55

Figure 5.1 BioMAG streptavidin beads with 0.01% Tween 20 in PBS (Image after

40seconds

Figure 5.2 BioMAG streptavidin beads with no surfactant in the supernatant (Image after 40seconds)

Magnet

Droplet periphery

Oil periphery

Magnet

Droplet periphery

Oil periphery

56

Figure 5.3 BioMAG streptavidin beads with 0.005% Tween 20 in PBS (Image after 40seconds)

Magnet

Droplet periphery

Oil periphery

57

5.3.2 Magnetic pull force

One of the most important parameters that would affect the efficiency of attraction with

very little aggregation is the magnetic field strength that is applied to the magnetically

responsive beads.

Materials- BioMag streptavidin beads suspended in 0.01 % Tween® 20, Dynal®

MyoneTM streptavidin beads suspended in 0.01% Tween® 20, 0.5 Tesla Neodymium

magnets (ND 42) with pull forces 5 lbs and 1.25 lbs, 1 Tesla Neodymium magnets with a

pull force of 1.25 lbs obtained from K & J Magnetics.

Experimental setup- BioMag streptavidin beads were diluted 4 times and suspended

in 0.01 % Tween® 20 and a 1µl droplet was sandwiched between Teflon coated

microscope slide and Teflon coated cover slip with 1.5 cSt silicone oil as the filler fluid.

The gap height of the sandwich used was 200 µm. A 0.5 Tesla Neodymium (ND 42)

magnet with a pull force of 5 lbs was placed at a distance of 5 mm from the bead droplet on

the cover slip and periodic images were taken using a microscope. The experiment was

repeated with a 0.5 Tesla magnet with a pull force of 1.25 lbs and a 1 Tesla magnet with a

pull force of 1.25 lbs.

Observations- It was observed that a magnet with a higher pull force (5 lbs) resulted

in aggregation of the beads and also adsorption to the surface of the top plate and the

microscope slide (Figure 5.4, Figure B.7). The 0.5 Tesla magnet (Figure 5.5) with 1.25 lbs

pull force provided efficient attraction with 50% less aggregation when compared to the

Figure 5.4. The magnetic strength/pull force of the magnet has to be chosen such that (1) it

is sufficiently strong to immobilize the magnetic beads (2) it is so strong that the beads

form aggregates/ clumps (3) it is not so strong that the resuspension occurs poorly when the

magnetic field is removed. For this reason we chose a 1Tesla magnet with a 1.25 lbs pull

force in this research.

58

Figure 5.4 BioMag streptavidin beads in 0.01 % Tween® 20 attracted with a 0.5 Tesla magnet (5 lbs pull force) (Image after 40 seconds)

Figure 5.5 BioMag streptavidin beads in 0.01 % Tween® 20 attracted with a 0.5 Tesla magnet (1.25 lbs pull force) (Image after 40 seconds)

Droplet periphery

Oil periphery

Oil periphery

Droplet

periphery

Magnet

Magnet

59

5.3.3 Concentration of the bead suspension

The solid content in the stock solution of the bead suspension is around 2-3%. The

concentration of the beads in a particular volume of the buffer is an important factor to

achieve high bead attraction efficiency. However concentration varies with size of the unit

droplet, which in turn depends on the electrode pitch and the gap height used.

Materials- BioMag Streptavidin beads suspended in 0.01 % Tween® 20, 1.5 cSt

Silicone oil

Experimental setup- The BioMag Streptavidin beads from the stock solution were

diluted 2, 4 and 6 times suspended in 0.01 % Tween 20. A 1 µl droplet of the beads was

sandwiched between a Teflon coated microscope slide and a cover slip with a gap height of

200 µm. The beads were attracted using a 1 Tesla magnet with a pull force of 1.25 lbs at a

distance of 5 mm from the droplet. The experiment was repeated with different

concentrations of the beads and periodic images were taken using a microscope.

Observations- It was observed that there was significant congestion and the

attraction was very poor for the highly concentrated beads (Figure 5.6, Figure B.6) when

compared to the 4 times diluted bead droplet (Figure 5.7). The attraction was more than

60% better in the case of the diluted bead droplet. However, the concentration of beads

used depends on the solid content of the bead suspension, which varies from stock to stock

and also on the electrode size, which determines the volume of the unit droplet.

5.3.4 Position of the magnet

The position of the magnet has a significant influence on bead attraction while

performing electrowetting operations to immobilize the beads and to remove the excess

supernatant. The process of washing magnetic beads basically involves repetition of the

bead droplet merging with a wash buffer solution, splitting and removal of the excess

supernatant and resuspension of the beads until acceptablmae levels of washing are

achieved. To achieve efficient washing, the magnet should be positioned strategically to

allow immobilization of the beads at a point and remove the excess supernatant with

minimal loss of beads.

60

Figure 5.6 BioMag Streptavidin beads (undiluted stock) in 0.01% Tween® 20 (Image after 40 seconds)

Figure 5.7 BioMag Streptavidin beads (4 times diluted) in 0.01% Tween® 20 (Image after 40 seconds)

Oil periphery

Droplet

periphery

Droplet

periphery

Oil

periphery

Magnet

Magnet

61

Magnetic field lines were simulated using Maxwell® software by Ansoft (Anonymous et

al., 2007h) with 1 Tesla neodymium magnets placed at different positions as described

below.

Configuration 1- Figure 5.8 shows a side view of the bead droplet sandwiched

between the chip with electrodes and a top plate with a spacing filled with 1.5 cSt Silicone

oil which is the filler fluid (not shown). This position of the magnet described in Figure 5.8

was simulated using Maxwell® software such that the “N” pole of the magnet was exactly

underneath the bead droplet. The simulation (Figure 5.9 shows a top view of the electrode

(small red rectangle) and the magnet underneath (large green rectangle) with the flow

happening from the top of the rectangle (start) to the bottom (splitting). The magnet is

bigger than the electrode on the chip and hence it covers more than one electrode. It can be

seen from the simulation that high surface field lines pass exactly through the center of the

electrode and gets weaker at the bottom which is the splitting zone to remove the excess

supernatant from the beads.

Configuration 2- In another configuration that is simulated for efficiently washing

the magnetic beads, the 1 Tesla magnets were arranged such that the opposite poles of both

the magnets face each other as depicted in the Figure 5.10. Magnets are positioned relative

to one or more transport electrodes in order to simulate the path of the magnetic field lines

through the liquid within the droplet. Figure 5.11 shows a top view of the magnets (large

rectangle) and the electrode (small rectangle). The software doesn’t allow overlapping of

the magnets; hence they were placed such that there is no overlap. It was observed that the

magnetic field lines pass through the edges of the electrodes from the north pole of the

magnet underneath to the south pole of the magnet over the electrode. This would form a

pillar of the beads as shown in Figure 5.10 which would enable lesser force for the splitting

of the supernatant.

62

Figure 5.8 1 Tesla magnet placed underneath the electrode with the bead droplet

Figure 5.9 Simulation of a 1 Tesla magnet placed under a magnetic bead droplet

Magnet Electrodes

Beads

Top plate

63

Figure 5.10 1 Tesla magnets placed underneath and over the bead droplets with the opposite poles facing each other

Figure 5.11 Simulation of 1 Tesla magnets placed underneath and over a bead droplet

Electrodes

Beads

64

Configuration 3- In another configuration simulated to efficiently wash the

magnetic beads, the magnets were arranged on the top and bottom of the bead droplet with

the opposite poles of the magnet facing each other and also magnets on either side of the

droplet with the poles facing each other as depicted in Figure 5.12. This ensures that the

magnetic field lines pass exactly through the droplet of beads which would retain them in

the place. The simulation (Figure 5.13) was done in Maxwell® software and it was

observed that the magnetic field lines were such that the beads will be retained exactly at

the center of the droplet. This ensures very less force required for the splitting process to

remove the excess supernatant. This type of configuration is typically called quadrapole

magnet arrangement.

Apart from the above mentioned magnet configurations, magnets can be placed in

any arrangement which can efficiently retain the beads in the droplet and ensure efficient

splitting of the supernatant. It should be mentioned that the splitting and the bead retention

should be reproducible. In all the configurations described above, the magnets can either be

permanent neodymium magnets or electromagnets. Permanent Neodymium magnets (ND

40 and ND 42) with surface field of 1 Tesla and 1.25 Tesla were used for this research. The

magnet configuration shown in Figure 5.8 with a magnet right underneath the electrode has

been used in this research because in all the other configurations, visualization would be

difficult with the magnets blocking the scope. Hence a 1 Tesla neodymium magnet (ND

42) was used for this research placed right underneath the electrode with the bead droplet.

The size of the ND 40 magnet used in this research was 0.2”X0.1”X0.048” which would

span at least three electrodes each spanning 1250 µm X 1250µm. Having magnetically

responsive beads retained at a centralized location within the droplet, the splitting operation

should occur such that the splitting zone is outside the surface field of the magnet (s).

Experiments were performed with the splitting zone within the magnetic field and

outside the magnetic field to quantify the loss of beads in both the operations.

65

Figure 5.12 1 Tesla magnets placed on four sides of the droplet

Figure 5.13 Simulation of the magnetic field lines in a quadrapole arrangement

Electrode

Beads

66

Materials- Dynal® MyoneTM Streptavidin coated beads (1.05 µm diameter), 0.01%

Tween® 20 in PBS, Parylene and 1% Teflon coated glass chip with 1250 µm X 1250 µm

chrome electrodes, 1% Teflon coated Indium Tin oxide (ITO) glass plate used as a top

plate for grounding the droplet.

Experimental setup- 1 µl of the Dynal streptavidin coated magnetically responsive

beads diluted 4 times from the stock solution was pipetted on an electrode and sandwiched

with the Teflon coated ITO top plate with the filler fluid being 1.5 cSt silicone oil. The gap

height in this case was 300 µm. 4 µl of 0.01 % Tween® 20 was added to the beads and

from the side of the top plate and allowed to settle for the beads to get attracted. The

splitting mechanism was done as shown in the Figures 5.14 and 5.15. The electrodes

adjacent to the one with the beads were switched on to form a slug of liquid in the first

stage. The electrodes at the end of the magnet where the surface field lines are weak were

switched off for the split to occur at that point. This retains the beads within the droplet and

also removes most of the excess supernatant. This left a droplet with all the magnetic beads

immobilized with the supernatant removed.

Observations- Figure 5.17 depicts the retention of the beads and splitting of the

droplet. However if the splitting happens within the zone where the magnetic field is strong

enough there would be loss of beads which in turn would diminish the signal. Figure 5.16

shows the loss of beads into the supernatant when the splitting happens in the vicinity of

the magnetic field. However when the droplet was split away from the magnetic field all

the beads were retained within the droplet which is shown in the Figure 5.17.

67

Figure 5.14 Splitting mechanism to retain the beads and remove the supernatant

Figure 5.15 Splitting of the supernatant retaining the beads

ON ON ON ON

1 Tesla magnet

ON OFF OFF ON ON

1 Tesla magnet

ON

ON

Stage 1

Stage 2

Splitting zone

68

Figure 5.16 Loss of beads with the splitting occurring within the effect of the magnetic field

1 Tesla

magnet

Electrodes

Loss of beads at

splitting zone

Beads lost after

split

Fewer beads due to loss at splitting

69

Figure 5.17 Splitting away from the affect of magnetic field

1 Tesla

magnet

Beads retained

Beads retained in the

droplet

Split away from

magnetic field

Bead free

supernatant

70

5.4 Washing of the magnetic beads

Washing of magnetic beads in an immunoassay involves repetitive droplet merging (with a

wash buffer solution), bead immobilization, splitting of the supernatant and bead

resuspension operations until acceptable levels of washing is achieved. Among the above

mentioned operations the bead retention and dilution efficiency are considered the most

important operations to achieve the acceptable wash efficiency retaining all or most of the

beads. Hence separate experiments were designed and performed to quantify the bead

retention and the dilution efficiency on chip.

5.4.1 Bead Retention

Materials- Dynal® MyoneTM Streptavidin coated magnetic beads (1.05 µm

diameter), Biotinylated Horse radish peroxidase (HRP) obtained from EY laboratories,

0.01% Tween® 20, Lumigen Ultra PS-Atto solutions A and B from Lumigen inc., Parylene

and 1% Teflon coated glass chip with 1250 µm X 1250 µm chrome electrodes, 1% Teflon

coated Indium Tin oxide (ITO) glass plate used as a top plate for grounding the droplet.

Experimental setup and Methods- Dynal® MyoneTM streptavidin coated magnetic

beads were labeled with HRP enzyme by reaction of the beads with the biotinylated HRP.

The stock concentration of the magnetic beads was 10mg/mL which can bind 20-25 µg of

biotinylated IgG. The stock concentration of the beads was diluted 4 times to enable

minimal loss of the beads at the splitting zone due to excess number of beads per unit

volume. 1 µl of the bead solution was added to 10 µl of 0.1 mg/ml of biotinylated HRP and

allowed to incubate for 30 minutes. The biotin on the HRP reacts with the streptavidin on

the magnetic beads to form a strong bond (K= 10^15 M) labeling the magnetic beads with

HRP. 1 µl of these HRP labeled beads were pipetted on an electrode and was sandwiched

using Teflon coated ITO top plate. The gap height used was 300 µm and was filled 1.5 cSt

Silicone oil with 0.1% Triton X 15. A 1 Tesla neodymium magnet placed underneath the

bead droplet and washing was done by adding 4 µl of 0.01 % Tween 20 for each wash as

described in the previous section.

71

This process was repeated 5 times and the supernatant was collected each time. The

supernatant should not have any beads lost. However in order to quantify the loss of beads

the supernatant was assayed in a Costar opaque 96 well plate. The supernatant was

collected from the chip and transferred into one of the wells in the plate. 5 µl of the

magnetic beads labeled with HRP was pipetted into another well and 5 µl of the Tween®

20 used for washing the bead was pipetted into another well of the 96 well plate. The

substrate was prepared by mixing equal volumes of Lumigen Ultra PS Atto solution A and

solution B and 50 µl of the substrate was added to each well and the chemiluminescence

was read using a BioTek plate reader for 8 minutes and the reading taken every 20 seconds

at a gain of 80. The mean velocity of the reaction was calculated by taking the slope of the

first five points for all the wells.

Mean V obtained for Tween 20 (negative control) = 900 mLum/min

Mean V obtained for the supernatant (data) = 1350 mLum/min

The quantification of the concentration of beads that were lost during the wash step can be

done using a standard curve between chemiluminescence values (mean V) and different

concentrations of HRP labeled beads.

Standard curve for HRP labeled magnetic beads- 2mg/ml of the HRP labeled beads

were diluted serially with Tween® 20 until 2E-9 mg/ml and 5 µl of each concentration of

the beads were transferred into a well of a 96 well Costar opaque plate. 50 µl of the

Lumigen Ultra PS atto substrate was added and the chemiluminescence read for 8 minutes

every 20 seconds.

Quantification of beads lost during the washing process-

Mean V value for the supernatant = 1350 mLum/min

Concentration range of the beads for the above mean V value = 2X10-7 to 2X10-6 mg/mL

No. of beads per mg of beads (stock solution) = 7-12 X 109 beads

No. of beads lost = 5X10-3 mL X 2X10-7 mg/mL

72

= 1X10-9 mg = 1X10-9X12X10-9

= 120 beads

Initial no. of beads in the 1 µl droplet = 25 E+6 beads

However, given the variability in the initial number of beads, the bead retention can

be approximated to be 100%.

Observation- Hence with the depicted magnet configuration and the splitting protocol

almost all the beads are retained within the droplet after removal of the excess supernatant.

5.4.2 Dilution efficiency

Another important operation to achieve good wash efficiency apart from good bead

retention is the dilution efficiency after every wash. An experiment was designed to

compare the wash dilution efficiency done on chip and with that obtained via conventional

laboratory analysis.

Materials- Dynal® Myone TM Streptavidin coated magnetic beads, Horse radish

peroxidase enzyme, 0.01 % Tween® 20, Parylene and 1% Teflon coated glass chip with

1250 µm X 1250 µm chrome electrodes, 1% Teflon coated Indium Tin oxide (ITO) glass

plate used as a top plate for grounding the droplet, Amplex red substrate for HRP enzyme.

Methods- Dynal streptavidin coated magnetic beads (1.05 µm diameter) were

diluted 4 times and suspended in 10 µg/ml of HRP enzyme prepared in 0.01 % Tween 20. 1

µl of these HRP suspended streptavidin coated magnetic beads were pipetted on an

electrode and sandwiched using an ITO top plate and the spacing is filled with 1.5 cSt

silicone oil. The gap height used was 300 µm. The magnet configuration described in

Figure 5.9 used for the bead retention experiments was used here. The bead droplet was

washed with 4 µl samples of 0.01 % Tween 20 and the supernatant was collected after

every wash and transferred to a well in a 96 well transparent plate. The same experiment

was performed even on bench with the washing done with a 96 well magnet holder and a

pipette to remove the supernatant. Amplex red substrate was added to each and every well

73

which had the supernatant after each wash from the bench and the chip. The absorbance

was read at a wavelength of 570 nm using the Biotek plate reader for 8 minutes. The mean

velocity was obtained by calculating the slope considering the initial 5 points.

Results- It was observed that the washing on the chip and in the laboratory bench

scale followed a similar trend with the supernatant (HRP enzyme) getting diluted after each

wash until the background absorbance was obtained (Figures 5.18 and 5.19). The above

experiment was done with different starting concentrations of HRP and different number of

wash steps was required to get the background absorbance values. The dilution curves on

the laboratory bench scale and chip were comparable which can be seen from Figure 5.19.

Table 5.1 shows the Mean V values of the reaction between the supernatant and the

substrate. Thus, the washing of the magnetically responsive beads was demonstrated on

chip using a 1 Tesla magnet right underneath the bead droplet with splitting of the

supernatant occurring away from the magnetic field. Table

5.5 Magnetic Immunoassay on chip

Washing of magnetically responsive beads which is the most important step to perform the

magnetic immunoassay on chip was described in the previous section. The magnetic

immunoassay was performed on chip considering all the parameters described above. The

analytes chosen to perform the magnetic immunoassay were Insulin and IL-6. The rationale

behind choosing insulin as the analyte is the prevalence of diabetes caused either due to

lack of insulin in the blood or because of development of insulin resistance by the cells.

5.5.1 Experimental setup

Materials- Access Ultrasensitive Insulin reagent pack (cat# 33410). The pack contains the

following reagents (Table 5.3). Lumigen APS-5 substrate (chemiluminescence substrate for

ALP from Lumigen Inc.), 1.5 cSt silicone oil, Access® wash buffer diluted 10 times using

0.05 M Tris and 0.1M NaCl, 12 µm Parylene and 6% Teflon coated glass chip (test chip

606) with 1250 µm X 1250 µm chrome electrodes, 1% Teflon coated Indium Tin oxide

(ITO) glass plate used as a top plate for grounding the droplet.

74

Figure 5.18 96 well plate comparing the washes on bench and chip

Comparison of washing

y = 1.0889x

R2 = 0.9892

0

50

100

150

200

250

300

0 50 100 150 200 250Washing done on bench scale equipment

Washin

g d

one o

n c

hip

Figure 5.19 Comparison of washing done on bench scale equipment and chip

On chip

On bench

75

Table 5.2 Mean V values of washes done on bench-scale equipment and on chip

Wash# Mean V (on chip) Mean V (on bench)

1 234 259.8

2 49.5 27.6

3 11.1 11.4

4 3 0.9

5 0.9 0.15

6 0 0

Table 5.3 Reagents in the Ultrasensitive Insulin reagent pack

R1a Mouse monoclonal anti-insulin coupled to paramagnetic particles,

TRIS buffer, bovine serum albumin (BSA matrix), <0.1% sodium

azide, and 0.1% ProClin 300

R1b Mouse monoclonal anti-insulin conjugated to bovine alkaline

phosphatase, TRIS buffer, BSA matrix, <0.1% sodium azide, and

ProClin 300.

R1c Mouse IgG in HEPES buffer, BSA matrix, <0.1% sodium azide, and

0.5% ProClin 300.

5.5.2 Chemiluminescence Detection

The whole chemiluminescence setup depicted in Figure 5.20 was set up in a dark box made

out of black foam board and aluminum tape. The darkness of the box was checked using

the PMT such that the background current obtained was around 3 nano Amperes. Since the

signals that are emitted from the reaction between the enzyme and the substrate are very

low even a small streak of stray light that might come from the LEDs on the camera can

cause a major error in the signal. Hence the box was built in a very robust manner such that

it is completely light tight.

76

Figure 5.20 Chemiluminescence detection setup

Teflon AF layer

Parylene layer

Chrome electrodes

Glass

Glass

ITO layer

Detector board

Photo multiplier tube

Droplet 1.5cSt Silicone oil H

Monitor

CCD camera

77

5.5.3 Experimental protocol on chip

The protocol depicted in Figure 5.21 was followed. 1 µl of the anti insulin antibody

coupled with the magnetically responsive beads was mixed with 1 µl of anti insulin

antibody labeled with ALP enzyme to form a 2 µl mixture in a tube. Since the ALP enzyme

has a tendency to adsorb to the magnetic beads and enhance the secondary signal, 1 µl of

blocking IgG was added to form a 3 µl of mixture. This mixture of the three antibodies was

pipetted on an electrode on the Test chip 606. 1 µl of a known concentration of insulin

sample was pipetted on another electrode in the same array on the chip. Both these droplets

were sandwiched using an ITO and Teflon coated top plate and the gap was filled with 1.5

cSt silicone oil (0.1% w/w Triton X 15). A gap height of 300 µm was maintained using a

shim. The droplets were brought closer to each other using electrowetting and merged and

allowed to incubate for 1 minute. Washing of the magnetic beads was done with the

washing protocol developed earlier to remove the excess supernatant. 4 µl of wash buffer

was used to dilute the supernatant for each wash. The beads were washed 5 times with 4 µl

of wash buffer each time. Then the beads were re-suspended in the solution surrounding the

beads by agitating the droplet using electrowetting. This re-suspended the beads and

avoided any clumping. 2 µl of Lumigen APS-5 substrate for the alkaline phosphatase

enzyme was added and the chemiluminescence was measured for every 1 second using the

photomultiplier tube (PMT) set up in the dark box for a span of 4 minutes. Figure 5.22

shows kinetic curves for concentrations of 0, 7, 70, 350 and 1050 pmol/L. To obtain a

standard curve (Figure 5.23) of the signal versus the insulin concentrations, the area under

each curve was calculated for each concentration and plotted against the insulin

concentrations. The same experiment was repeated thrice on different days and the results

were reproducible with an error of less than 3%. Thus, a magnetic immunoassay on insulin

was performed on a digital microfluidic platform involving multiple droplet operations

such as merging, splitting, washing of magnetic beads which involved immobilization of

beads within the droplet and resuspension of beads and on chip detection. The assay was

reproducible with an error of less than 3% between different runs on different days. The

least concentration of insulin detectable was 0.24 pg/µL using the magnetic immunoassay.

The sensitivity can further be improved by reducing the distance between the PMT and the

droplet. However to validate the data obtained on chip, the same experiment was repeated.

78

Step 1 Step 2

Step 3Step 4

Magnetic

beads with

antibodies

Analyte

Attraction

using magnet

Splitting of

supernatant

Figure 5.21 Step by step protocol of magnetic immunoassay on chip

79

Figure 5.22 Kinetic curves of different concentrations of insulin (magnetic immunoassay)

Figure 5.23 Insulin standard curve on chip

Insulin standard curve on chip

y = 339291x + 1E+08

R2 = 0.944

0.0E+00

1.0E+08

2.0E+08

3.0E+08

4.0E+08

5.0E+08

6.0E+08

0 200 400 600 800 1000 1200

Insulin (pmol/L)

Are

a u

nd

er

the

cu

rve

,

ch

em

ilu

min

esce

nce

Kinetic curves of insulin on chip

0.00E+00

5.00E+05

1.00E+06

1.50E+06

2.00E+06

2.50E+06

0 50 100 150 200 250

Time, seconds

AD

C c

ounts

,

chem

ilum

inescence

S0- 0 pmol/L

S1- 7 pmol/L

S2- 70 pmol/L

S3- 350 pmol/L

S4- 1050 pmol/L

80

on conventional bench scale equipment. The washing was performed by immobilizing the

beads using a Neodymium magnet and removing the supernatant using a pipette.

5.5.4 Experimental protocol on conventional bench scale equipment

The same experiment described in section 5.5.3 was repeated on bench in tubes to validate

the data obtained on chip. 1 µl of anti-insulin antibody labeled with magnetic beads was

mixed with 1 µl of anti-insulin antibody labeled with ALP enzyme to which 1 µl of

blocking IgG was added. A 1 µl sample of a known concentration of insulin was added to

the above mixture and allowed to incubate for 5 minutes in the tube. The excess/unreacted

supernatant was later removed by immobilizing the magnetic beads using a 1 Tesla magnet.

The beads were then washed 5 times with 4 µl sample of wash buffer each time. Later the

beads were resuspended in 1 µl of wash buffer and 2 µl of Lumigen APS-5 substrate for the

ALP enzyme was added. Chemiluminescence from the reaction was measured using the

PMT under the same conditions that were done on chip. The chemiluminescence readings

were measured every second for a span of 4 minutes and similar analysis was done as

described above. The areas under each kinetic curve (Figure 5.24) were calculated and

plotted against corresponding concentration of insulin to obtain the standard curve shown

in Figure 5.25. The slope of the straight line obtained was 6.35X106 with an R2 value of

0.9977. The first four concentrations on the bench and the chip matched perfectly.

However the signal for the highest concentration of insulin (1050 pmol/L) on chip was

different from that obtained using conventional bench scale equipment which is because of

lesser amount of substrate (Lumigen APS-5) used. A plot between the data obtained on

conventional bench scale equipment and data on-chip was drawn (Figure 5.26). The

straight line had an intercept of 0 with a slope of 0.9513 which confirms that the assays

done on conventional bench scale equipment and on chip were comparable with 95%

confidence levels. The background obtained for the zero concentration of insulin on

conventional bench scale equipment was almost similar to that obtained on chip which

reinstates that the washing on chip was very good. Hence a magnetic immunoassay with a

detection sensitivity of 0.24 pg/µL was performed on a digital microfluidic platform.

81

Kinetic curves of insulin on bench

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

2.5E+06

3.0E+06

3.5E+06

4.0E+06

4.5E+06

0 50 100 150 200 250

Time (seconds)

AD

C c

ou

nts

, C

he

milu

min

esce

nce

S4-1050 pmol/L

S3-350 pmol/L

S2-70 pmol/L

S1-7 pmol/L

S0-0 pmol/L

Figure 5.24 Kinetic curves for different concentrations of insulin on bench-scale equipment

Figure 5.25 Insulin standard curve on bench-scale equipment

Insulin standard curve on bench

y = 634759x + 8E+07

R2 = 0.9977

0.0E+00

1.0E+08

2.0E+08

3.0E+08

4.0E+08

5.0E+08

6.0E+08

7.0E+08

8.0E+08

0 200 400 600 800 1000 1200

Insulin (pmol/L)

Are

a u

nd

er

the

cu

rve

,

ch

em

ilu

min

esce

nce

82

y = 0.9513x

R2 = 0.9775

0.00E+00

5.00E+07

1.00E+08

1.50E+08

2.00E+08

2.50E+08

3.00E+08

3.50E+08

0.00E+00 5.00E+07 1.00E+08 1.50E+08 2.00E+08 2.50E+08 3.00E+08 3.50E+08

On conventional bench scale equipment

On c

hip

Figure 5.26 Comparison of Insulin assay done on bench-scale equipment and chip

83

5.6 Insulin assay on serum

Materials-

• Serum with an unknown concentration of insulin from a non-diabetes patient was

obtained from Sigma,

• Insulin standard (1050 pmol/L) from Beckmann Coulter Inc.,

• Access Immunoassay kit from Beckmann Coulter Inc.

• Lumigen APS-5 substrate for alkaline phosphatase,

• 1.5 cSt Silicone oil (0.1% w/w Triton X 15),

• Test chip 606 coated with 12 µm parylene and coated with 6% Teflon,

• ITO coated glass plate coated with 1% Teflon which acts as a ground electrode.

Methods- Aliquots of serum were thawed and doped with insulin by mixing 1 part of

insulin standard (1050 pmol/L) with 4 parts of serum and 1 part of insulin standard (1050

pmol/L) with 9 parts of serum. The complete immunoassay was performed with 4 samples

which included 0 pmol/L concentration of insulin. The chemiluminescence was collected

with the help of a PMT placed right over the droplet and measured in terms of ADC counts.

Inference- It can be observed from the kinetic curves (Figure 5.27) that the

chemiluminescent signal increased with the increase in insulin concentration. The insulin

concentration in the serum was calculated by utilizing the standard curve on chip obtained

earlier and it was found that the serum had 34.65 pmol/L of insulin concentration. The

experiment was repeated on the bench and the insulin concentration obtained was 35.61

pmol/L. The kinetic curves for the immunoassay for insulin done on serum, on-chip and on

conventional bench-scale equipment are shown in Figures 5.28 and 5.29 respectively. The

concentration of insulin in the serum of a healthy individual ranges from 30-45 pmol/L.

The concentrations of insulin that were obtained from the immunoassays performed

compared well with the typical non-diabetic individual’s serum insulin concentration.

84

Figure 5.27 Kinetic curves of magnetic immunoassay on serum for Insulin

Insulin assay on serum

3.0E+05

4.0E+05

5.0E+05

6.0E+05

7.0E+05

8.0E+05

9.0E+05

0 50 100 150 200 250

Time (seconds)

AD

C c

ounts

, chem

ilum

inescece

1 part Insulin + 4

parts serum

1 part insulin + 9

parts serum

serum alone

(insulin

concentration

unknown)S0 - 0 pmol/L

85

Figure 5.28 Kinetic curve of Insulin assay on serum on bench-scale equipment

Figure 5.29 Kinetic curve of Insulin assay on serum on chip

Insulin assay on serum on chip

395000

405000

415000

425000

435000

445000

455000

465000

475000

485000

495000

505000

0 50 100 150 200 250

Time (seconds)

Ch

em

ilu

min

es

ce

nc

e,

AD

C c

ou

nts

Insulin assay on serum on bench

395000

400000

405000

410000

415000

420000

425000

430000

435000

440000

445000

450000

0 50 100 150 200 250

Time (seconds)

Ch

em

ilu

min

esce

nce

(A

DC

co

un

ts)

86

5.7 Magnetic immunoassay of Interleukin-6 (IL-6) on chip

5.7.1 Immunoassay of IL-6 on digital microfluidic chip

Materials- Access® IL-6 reagent pack (cat# A 30945). The pack contains the following

reagents (Table 5.4), Lumigen APS-5 substrate from Lumigen Inc. for alkaline

phosphatase, 1.5 cSt silicone oil with 0.1% Triton X-15, Glass chip coated with 12 µm

parylene and dip coated with 6% Teflon, ITO coated glass plate dip coated with 1% Teflon

which acts as the ground plane.

Table 5.4 Reagents in Access® Immunoassay IL-6 kit

R1a

Paramagnetic particles coated with goat anti-mouse IgG:mouse

antihuman IL-6 monoclonal antibody, BSA, surfactant, 0.1%

sodium azide and 0.17% Proclin 300

R1b

Tris saline buffer, proteins (porcine, goat, bovine, mouse),

surfactant, 0.1% sodium azide and 0.17% Proclin 300

R1c

Goat anti-human IL-6 alkaline phosphatase(bovine) conjugate,

BSA, surfactant, 0.1% sodium azide and 0.17% Proclin 300

Methods- The immunoassay on the glass chip was performed following the same

protocol as depicted in Figure 5.22. 1µL each of R1a, R1b and R1c were pipetted manually

on to an electrode on the chip and 1 µL of the IL-6 sample was pipetted at different

electrode on the same electrode array. Both these droplets were sandwiched by placing a

ITO coated glass plate over the droplets which acts as the ground plane. A gap height of

300µm was maintained which was filled with 1.5 cSt silicon oil with 0.1% Triton® X 15.

Both the droplets were merged using electrowetting and allowed to incubate for 2 minutes

with agitation performed by transporting over 4-5 electrodes. The whole reaction mixture

was taken to the electrode with the 1 Tesla magnet underneath and the supernatant was

removed using the washing protocol developed earlier in the chapter. Washing was

87

performed repetitively until no signal was seen in the supernatant using 4 µL wash samples

for each wash. 2 µL of the substrate (Lumigen APS-5) was added to the washed magnetic

beads and the chemiluminescence was collected with a PMT placed right over the droplet

for 4 minutes in terms of ADC counts.

Analysis- The kinetic curves of chemiluminescence were plotted in terms of ADC

counts versus time in seconds shown in Figure 5.30. The area under each kinetic curve was

calculated by integrating the curve for 240 seconds and the area for each concentration was

plotted against the concentration to obtain a standard curve on chip as shown in Figure

5.31. It was calculated that the lowest concentration of IL-6 that can be detected using this

droplet based lab-on-a-chip magnetic immunoassay is 0.24 pg/µL. The data was

reproducible with a standard error of less than 2 %.

5.7.2 Immunoassay of IL-6 on conventional bench-scale equipment

The same experiment described in section 5.7.1 was repeated on conventional bench scale

equipment in tubes to validate the data obtained on chip. 1 µl of anti-IL-6 antibody labeled

with magnetic beads was mixed with 1 µl of anti-IL-6 antibody labeled with ALP enzyme

to which 1 µl of block IgG was added. 1 µl of the IL-6 samples of different concentrations

(0, 2.5, 25, 250, 750 pg/mL) was added to different tubes with the antibodies and the

magnetic beads and allowed to incubate for 5 minutes. The excess/unreacted supernatant

was later removed by immobilizing the magnetic beads using a 1 Tesla magnet. The beads

were then washed 5 times with 4 µl sample of wash buffer each time. Later the beads were

resuspended in 1 µl of wash buffer and a 2 µl of Lumigen APS-5 chemiluminescent

substrate for the ALP enzyme was added and the chemiluminescence was measured under

the same conditions as the chip. The chemiluminescence readings were taken every second

for a span of 4 minutes and similar analysis was done as described above. The area under

each kinetic curve was calculated and plotted against the corresponding concentration of

insulin to obtain the following standard curve (Figure 5.32). The slope of the straight line

obtained was 1.18X106 with an R2 value of 0.987.

88

Figure 5.30 Kinetic curves of IL-6 immunoassay on lab-on-a-chip

Figure 5.31 Standard curve of IL-6 immunoassay on lab-on-a-chip

Kinetic curves of IL-6 on chip

0.00E+00

1.00E+06

2.00E+06

3.00E+06

4.00E+06

5.00E+06

0 50 100 150 200 250

Time (seconds)

Chem

ilum

inescence,

AD

C c

ounts

S4-750 pg/mL

S3- 250 pg/mL

S2- 25 pg/mL

S1- 2.5pg/mL

S0- 0 pg/mL

IL-6 standard curve on chip

y = 1.18E+06x + 8.80E+07

R2 = 9.87E-01

0.00E+00

2.00E+08

4.00E+08

6.00E+08

8.00E+08

1.00E+09

1.20E+09

0 100 200 300 400 500 600 700 800

IL-6 (pg/mL)

Are

a u

nd

er

the

cu

rve

89

Figure 5.32 IL-6 standard curve on bench

y = 1.0572x

R2 = 0.9916

0.00E+00

2.00E+08

4.00E+08

6.00E+08

8.00E+08

1.00E+09

1.20E+09

0 2E+08 4E+08 6E+08 8E+08 1E+09

On conventional bench scale equipment

On

-ch

ip

Figure 5.33 Comparison of IL-6 immunoassay on bench and chip

IL-6 standard curve on bench

y = 1.11E+06x + 8.80E+07

R2 = 9.95E-01

0.00E+00

2.00E+08

4.00E+08

6.00E+08

8.00E+08

1.00E+09

1.20E+09

0 200 400 600 800

IL-6 (pg/mL)

Are

a u

nd

er

the

cu

rve

90

A plot between the areas obtained from the assay done on conventional bench scale

equipment and on chip was drawn (Figure 5.33). The straight line passed through zero with

a slope of 1.057 and an R2 value of 0.9916. Thus, it can be seen that the data on chip and

done on conventional bench scale equipment were very comparable.

5.8 Immunoassay of IL-6 on serum

Materials- Serum with an unknown concentration of IL-6 obtained from Sigma, anti-IL-6

antibody coupled with magnetic beads (reagent R1a), anti-IL-6 antibody labeled with

alkaline phosphatase enzyme (reagent R1c) and blocking antibody (reagent R1b) from the

Access® immunoassay kit from Beckmann Coulter Inc., Access wash buffer diluted 10

times by a buffer with 0.5M Tris and 0.01 M Nacl, Lumigen APS-5 substrate for alkaline

phosphatase, 1.5 cSt Silicone oil, Test chip 606 coated with 12 µm parylene C and coated

with 6% Teflon, ITO coated glass plate coated with 1% Teflon which acts as a ground

electrode. The chemiluminescence was measured with the help of a PMT placed right over

the droplet and measured in terms of ADC counts. The kinetic curves obtained for the

assay done on serum done on conventional bench-scale equipment and the chips were

comparable with 95% confidence levels. It was seen that the assay done on bench gave a

slightly higher signal which is attributed to longer incubation of the samples.

Methods- 1 µL of magnetic beads labeled with anti-IL-6 antibody was mixed with 1

µL of anti-IL-6 antibody labeled with ALP followed by block IgG to form a 3 µL reagent

mixture. 1 µL of serum containing unknown concentration of IL-6 was added to the above

reagent mixture and allowed to incubate for 2 minutes. The supernatant was removed and

the magnetic beads were washed using the protocol developed. 1 µL of Lumigen APS-5

was added to the washed magnetic beads and the chemiluminescence was measured for 4

minutes using a PMT placed right over the magnetic bead droplet. The assay was repeated

using conventional bench scale equipment.

Analysis- The areas under the kinetic curves were calculated and the concentrations

of IL-6 were calculated from the respective standard curves (Figures 5.34 and 5.35).

91

Figure 5.34 Kinetic curve of IL-6 assay on serum on chip

Figure 5.35 Kinetic curve of IL-6 assay on serum on bench

IL-6 assay on serum on chip

8.00E+05

9.00E+05

1.00E+06

1.10E+06

1.20E+06

0 50 100 150 200 250

Time, seconds

Chem

ilum

inescence (A

DC

counts

)

IL-6 assay on serum on bench

8.00E+05

9.00E+05

1.00E+06

1.10E+06

1.20E+06

0 50 100 150 200 250

Time (seconds)

AD

C c

ounts

, chem

ilum

inescence

92

The concentration of IL-6 obtained from the bench assay was 140 pg/mL and that from the

chip was 126 pg/mL. The data was reproducible with a standard error of 2% on different

days and different chips.

5.9 Cross Contamination of the analytes with the magnetic beads

Human serum contains all the analytes including insulin and IL-6. If one is testing for one

particular analyte in human serum or blood or other physiological fluid, there should be no

secondary reaction and contamination from the other analytes. Hence the following

experiments were performed to establish that there is no secondary reaction of the analytes

with the magnetic beads

Materials-

• Access® Insulin and IL-6 Immunoassay kits from Beckmann Coulter Inc.,

• Lumigen APS-5 substrate for Alkaline phosphatase,

• Test chip 606 coated with 12µm parylene and 6 % Teflon

• ITO top plate coated with 1% Teflon which acts as the ground plane

Methods-

Experiment 1- 1 µL droplet of IL-6 (250 pg/mL) was mixed with 1 µL of anti-IL-6

antibody labeled with Alkaline phosphatase and is merged with a 1 µL droplet of magnetic

beads labeled with anti-Insulin antibody. The reagent mixture was allowed to incubate for 2

minutes and the supernatant was removed and the magnetic beads were washed using the

protocol developed. 1 µL of Lumigen APS-5 was added to the washed magnetic beads and

the chemiluminescence was measured for 4 minutes using a PMT placed right over the

magnetic bead droplet. The area under the curve was calculated and it was obtained to be

84.5 E6 which is equal to the background. This establishes that there is no secondary

reaction between the IL-6 antigen and the magnetic beads.

93

Experiment 2- 1 µL droplet of Insulin (350 pmol/L) was mixed with 1 µL of anti-Insulin

antibody labeled with Alkaline phosphatase and is merged with a 1 µL droplet of magnetic

beads labeled with anti-IL-6 antibody. The reagent mixture was allowed to incubate for 2

minutes and the supernatant was removed and the magnetic beads were washed using the

protocol developed. 1 µL of Lumigen APS-5 was added to the washed magnetic beads and

the chemiluminescence was measured for 4 minutes using a PMT placed right over the

magnetic bead droplet. The area under the curve was calculated and it was obtained to be

89.2 E6 which is equal to the background. This establishes that there is no secondary

reaction between the Insulin antigen and the magnetic beads.

The above result establishes that different analytes present in the serum would not affect

the result of the analyte that is being assayed due to secondary adsorption onto magnetic

beads.

5.9 Chapter Summary

A protocol for effectively washing the magnetic beads on droplet based lab-on-a-chip was

developed. Optimal strength, position of the magnet was simulated and experimentally

determined to effectively immobilize all the magnetically responsive beads without

aggregation and resuspend substantially all the beads after washing. The washing protocol

developed provided bead retention of almost 100%. Magnetic immunoassays were

performed on insulin and IL-6 on this droplet based lab-on-a-chip utilizing the washing

protocol developed. The sensitivities achieved were 0.24 pico grams/µL for insulin and 4

femto grams/µL for IL-6. The data obtained on chip was comparable with the data obtained

using conventional bench-scale equipment. The results were reproducible with <3%

standard error between different runs done on different chips on different days.

94

CHAPTER 6

CONCLUSIONS AND FUTURE WORK

6.1 Conclusions

A programmable digital microfluidic lab-on-a-chip to perform immunoassays based on

magnetically responsive beads was developed. An efficient protocol for washing of the

magnetic beads was developed. Immunoassays for insulin and IL-6 were performed on the

lab-on-a-chip based on digital microfluidics.

6.1.1 Lab-on-a-chip fabrication and testing

A simple architecture for the lab-on-a-chip, integrating the previously developed digital

microfluidic components was designed to perform and show the proof of concept of a

magnetic immunoassay on chip. The major concern in the fabrication was the thickness of

the insulator which caused electrolysis. This reduced the time of duration of each

experiment to approximately 10 minutes. Though the exact reason for the breakdown is

unknown at this point of time, it is surmised that it might due to mechanical cracking or

failure of the parylene film at the gasket-electrode junction which exposes the metal to the

liquid and causes electrolysis. However, a 12 µm parylene film was used which improved

the performance of the chips and significantly increased the life time of the chips. The

coating of the Teflon AF also had a significant affect on the life time of chips. It was

observed that a 6 % Teflon AF coated chip had a higher life time when compared to a 1%

Teflon dip coated chip. The biocompatibility of the droplet based digital microfluidic lab-

on-a-chip was established by transporting different proteins and surfactants that would be

used in the immunoassay. It was found that a unit droplet with magnetically responsive

beads can be easily transported on the digital microfluidic platform which was a major

concern for all the commercially available continuous microfluidic based lab-on-a-chips.

The magnetic beads easily immobilized and resuspended very well by transporting the

droplet across a span of five electrodes. Protein fouling on the droplet based microfluidic

95

chips was confirmed and quantified by transporting a 1 µL droplet of 1 mg/mL HRP and

later analyzing by transporting a HRP substrate droplet. Hence new chips and new

electrode arrays were used for each separate experiment minimizing secondary reactions.

6.1.2 Washing of magnetic beads

All the parameters involved in washing of magnetically responsive beads on droplet based

digital microfluidic lab-on-a-chip were studied and optimum conditions to achieve efficient

washing were found out. It was found out that the beads require an optimum percentage of

surfactant in the buffer surrounding the magnetic beads. The percentage of

surfactant for the magnetically responsive beads to attract efficiently and resuspend

completely after the removal of the magnetic field without formation of any clumps and

aggregates was found to be 0.01% (weight/weight). Various neodymium permanent

magnets with a wide range of surface fields and pull forces were studied to attract the

magnetic beads. It was found out that neodymium magnets with a higher surface field and

lesser pull force provided efficient attraction with minimum clumping of the magnetically

responsive beads. A 1 Tesla magnet with a pull force of 1.25 lbs was chosen for attracting

the magnetic beads in this thesis which provided efficient attraction with minimum

aggregation and also enabled substantial resuspension of the beads. Different

configurations of the magnet positions were simulated using Maxwell® software and the

magnetic field lines were analyzed. The different configurations simulated were

1. 1 Tesla neodymium magnet placed right under the droplet containing magnetic

beads

2. 1 Tesla neodymium magnets placed right over and under the droplet with the

magnetic beads with the opposite poles facing each other.

3. 1 Tesla neodymium magnets placed on the four sides of the droplets with the

opposite poles facing each other (also termed as quadrapole magnetic arrangement)

The magnetic configuration utilized in this thesis was Configuration#1 to ensure recording

of the operations that are being carried out on the lab-on-a-chip. A stage was built using

acrylic with grooves that exactly fit the 1 Tesla magnets and align with the electrode arrays.

Washing is done by attracting or immobilizing the beads using the 1 Tesla magnet placed

underneath and removing the excess supernatant by creating a slug using electrowetting

96

and splitting off at a point away from the magnetic field. This process was repeated until

the washing levels were achieved. This would ensure efficient washing and also significant

bead retention within the droplet. The point at which splitting occurred was away from the

magnetic field so that the beads were retained within the droplet and do not escape into the

supernatant through the neck at the point of split. Hence the washing protocol was

developed in this thesis in such a way as to

1. avoid permanent clumping and aggregation of the magnetically responsive beads

2. during a droplet splitting operation, capture and immobilize substantially all of the

magnetically responsive beads within a single droplet

3. ensure immobilization and retention of substantially all of the magnetically

responsive beads during the washing operation and

4. upon completion of washing process, ensure resuspension of substantially all of the

magnetically responsive beads within the liquid and with substantially no clumping

or aggregation thereof.

Utilizing the magnetic configuration explained in this section, washing of magnetic beads

suspended in 10 µg/mL of free HRP enzyme was performed and compared with the

washing done on bench manually. The results on bench and chip were very comparable to

each other and it required six repetitive washes with 5 µL wash samples to achieve

complete washing. It was found that the bead retention during the washing using the above

developed protocol was ~ 100%.

6.1.3 Magnetic immunoassays of Insulin and IL-6 on droplet based digital

microfluidic lab-on-a-chip

Utilizing the washing protocol developed, magnetic immunoassay was performed on

insulin analyte and a standard curve of insulin was achieved. The lowest amount of insulin

that was detectable on chip was 0.24 pg. It was very well comparable with the standard

curve obtained by performing the assay on bench under the same conditions. The above

immunoassay was performed using the samples prepared in buffer. The immunoassay was

also performed with serum alone and serum doped with different concentrations of insulin.

It was observed that the signal increased with increasing insulin concentration in the serum.

The insulin concentration that was determined in the serum using the standard curve

97

developed was found to be 34.61 pmol/L and that obtained from the assay performed on

bench was 35.61 pmol/L. The same assay was performed on IL-6 analyte and the lowest

amount that was detectable 0.4 fg. The standard curve compared very well with that of the

curve obtained on bench. The immunoassay was also performed on serum for IL-6 both on

bench and chip and they compared very well with very good reproducibility. The

concentration of IL-6 that was obtained from the assay on serum was 126 pg/mL and 140

pg/mL respectively on chip and bench. Hence immunoassays on two analytes (insulin and

IL-6) were performed on a droplet-based digital microfluidic lab-on-a-chip involving

magnetically responsive beads.

6.2 Future work

The developed lab-on-a-chip based on digital microfluidics establishes that an

immunoassay involving magnetically responsive beads can be performed on a chip. For the

developed lab-on-a-chip to be made into a commercial product, it is necessary to further

integrate and automate the detection instrumentation and address packaging issues.

6.2.1 Lab-on-a-chip architecture

The proposed lab-on-a-chip which can handle or perform multiple immunoassays using

magnetic beads in a completely automated fashion has reservoirs for all the reagents so that

the droplets of the primary antibody couple with beads, secondary antibodies, analyte from

serum or plasma or blood can be dispensed based on the protocol developed on bench and

washed with the wash buffer from another reservoir utilizing the wash protocol developed

in Chapter 5. After sufficient washing of the magnetic beads, the substrate is added and the

chemiluminescence is detected using the integrated detection instrumentation. Since the

wash protocol developed is automated, the automatic dispensing of proteins from the

reservoirs was already established by previous researchers, the development of a

completely automated immunoassay platform with magnetic beads can be achieved.

98

6.2.2 Detection methodologies

Chemiluminescent detection is the easiest to integrate with the electrowetting-based-lab-

on-a-chip platform. The lowest concentration of IL-6 that was detectable using the PMT

(Hamamatsu’s 9858) was 4 fg/µL. However to achieve more sensitivity in order to perform

single cell assays, a photon counting instrument would be orders of magnitude more

sensitive than the PMT. A photon counting unit works on the similar principle as a PMT

where in the incident light striking the photocathode is amplified by the cascade process of

secondary emission through the dynodes (normally 106 to 107 times) and finally reach the

anode connected to an output circuit. Other optical methods such as fluorescent methods of

detection require a complicated set up involving a light source and different combinations

of filters and absorbance is not as sensitive when compared to other detection strategies.

Electrochemical detection is a more suitable detection methodology but has the drawback

of adding to the fabrication complexity in the device. However electrochemical methods

are unavoidable if whole blood analysis is required at the places of point-of-care [Wang,

2002].

6.2.3 System integration

The major concern in any completely integrated and automated point-of-care lab-on-a-chip

device is efficient sample preparation methods. This is infact the biggest impediment to the

commercial acceptability of microfluidic technologies and considerable research is required

in this area [deMello, 2003]. The main issue in sample preparation would be the separation

of cells from whole blood to obtain serum or plasma on chip. Apart from this packaging of

all the small pieces into a single device would be critical since the device would be used in

widely varying environmental conditions. Furthermore the storage of all the antibodies

which contain proteins is another issue if these devices would be used in a clinical point-of-

care setting as disposable cartridges.

99

NOMENCLATURE

LV

SL

SV

θ

E

ε0

εr

V

d

Interfacial tension between liquid and filler medium, N/m

Interfacial tension between solid and liquid droplet, N/m

Interfacial tension between solid and filler medium, N/m

Contact angle, degrees

Energy stored in the parallel plate capacitor, J

Permittivity of vacuum, F/m

Relative permittivity of the insulator

Voltage applied, V

Thickness of insulator, µm

100

APPENDIX A

ANALYSIS OF CHEMILUMINESCENT DATA USING

ENZYME KINETICS

The enzyme-substrate reaction was proposed to be composed of two elementary reactions

in which the substrate forms a complex with the enzyme that subsequently decomposes to

products and enzyme:

E+S ES P+E

Here E, S, ES and P symbolize the enzyme, substrate, enzyme-substrate complex and the

product respectively (for enzymes composed of multiple identical subunits E refers to

active sites rather than enzyme molecules). The general expression for the velocity (rate) of

this reaction is

][][

2 ESkdt

Pd==ν A1.1

The overall rate of production of [ES] is the difference between the rates of elementary

reaction leading to its appearance and those leading to its disappearance

][][]][[][

211 ESkESkSEkdt

ESd−−= − A1.2

Assuming steady state,

0][=

dt

ESd A1.3

The quantities [E] and [ES] are not directly measurable but the total enzyme concentration,

][][][ ESEE T += A1.4

is usually readily determined. The rate equation is then derived as follows.

Combining Eq A1.2 with the steady state assumption and Eq A 1.4 yields:

])[(]])[[]([ 211 ESkkSESEk T +=− − A1.5

which upon rearrangement becomes

][][])[]([ 1111 SEkSkkkES T=++ −− A1.6

101

Dividing both sides by k1 and solving for [ES]

][

][][][

SK

SEES

M

T

+= A1.7

Where KM is known as the Michaelis constant defined as

1

21

k

kkK M

+= − A1.8

From the definition of the rate of the reaction,

A1.9

A1.10

The maximal velocity of a reaction Vmax, occurs at high substrate concentration ns when

the enzyme is saturated that is, when it is entirely in its [ES] form:

TEkV ][2max = A1.11

The Michaelis constant has a simple operational definition. At the substrate concentration

where [S] = KM,

ν0=Vmax/2, so that KM is the substrate concentration at which the reaction velocity is half-

maximum.

So in this research the substrate concentration used is way higher than the KM value which

ensures maximum velocity. The kinetic curve (eq A1.9) is integrated over time to obtain

the concentration of enzyme which is directly proportional to that of the analyte captured.

∫ +=

+===

t

M

T

M

T

dtSK

SEkP

SK

SEkESk

dt

Pdv

0

2

22

][

][][][

][

][][][

][

102

APPENDIX B

IMAGE ANALYSIS USING IMAGE J® SOFTWARE

Image analysis was done using Image J (Anonymous et al., 2007 i) software using

the following particle analysis. This information is obtained from the Image J manual. The

raw data was made changed using the thresholding and watershedding effects to analyze

individual particles as shown in the Figure below

Figure B.1 Analyzing particles using Image J

B1.1 Threshold segmentation

Automatic particle analysis requires the image to be a “binary” image i.e. black or white.

The software needs to know exactly where the edges are to perform morphology

measurements. A “threshold” range is set and pixels in the image whose value lies in this

range are converted to black; pixels with values outside this range are converted to white

(or vice versa depending on the user’s request).

B1.2 Watershed segmentation

The image first needs to be converted to a binary (via thresholding). The black pixels are

then replaced with grey pixels of an intensity proportional to their distance from the white

pixel (i.e. black pixels close to the edge are light grey, those closer to the middle are nearer

black). This is an Euclidian distance map (EDM). From this it calculates the centers of the

objects the ultimate eroded points (UEPs) i.e. points that are equidistant from the edges.

103

These points are then dilated until they meet another black pixel, then a water shed line is

drawn. This is shown in the Figure

Figure B.2 Watershed segmentation using Image J

B1.3 Analyze Particles

Once the image has been segmented, the menu command “Analyze/Analyze particles” can

be used to obtain various information regarding the particle numbers. Set the minimum size

and maximum size to exclude objects that appear in the binary image that are clearly not

objects of interest. The following images (Figures B.3, B.5 and B.4) show thresholded

images of the raw data shown in Figures 5.2, 5.3 and 5.4 respectively.

Figure B.3 Thresholded image of magnetic beads in 0.01% Tween 20 (raw data-Figure 5.2)

104

Figure B.4 Thresholded image of magnetic beads in 0.005% Tween 20 (raw data-Figure 5.4)

Figure B.5 Thresholded image of magnetic beads in 0% Tween 20 (raw data-Figure 5.3)

105

Figure B.6 Thresholded image of undiluted stock of magnetic beads (raw data-Figure 5.5)

Figure B.7 Thresholded image of beads attracted with a magnet of high pull force (raw data Figure-5.7)

106

BIBILIOGRAPHY

(1) Adams D., Frank M., “Point-of-care technology: The i-STAT system for bedside blood

analysis,” J. of Pediatric nursing, 1995; 10: p194

(2) Adamson A., Gast A., “Physical Chemistry of Surfaces,” John Wiley and Sons Inc., 6th

Edition, 1997

(3) Aller R., “Chemistry analyzers branching out,” CAP Today, 2002; p 84-106

(4) Anoymous, “LabChip®3000 Drug Discovery System,"

http://www.caliperls.com/products/labchip3000.html, March 2007a (5) Anonymous, “Gyros(R) Work station LIF,”

http://www.gyros.com/products/gyrolab_workstation_lif.html, March 2007b

(6) Anonymous, “Liquid handling and robotics," http://www.tecan.com/page/content/index.asp?MenuID=1&ID=2&Menu=1&Item=21.1, March 2007c

(7) Anonymous, “ActiveTM cards,” http://micronics.net/products/active.php, March 2007d

(8) Anonymous, “"Better outcomes begin with better diagnosis,"

http://www.biosite.com/products/default.aspx, March 2007e

(9) Anonymous, “"iSTAT-Product info," http://www.abbottpointofcare.com/istat/www/products/index.htm, March 2007f

(10) Anonymous, “Detectors,” http://sales.hamamatsu.com/en/products/electron-tube-

division/detectors/photomultiplier-modules/h9858.php, March 2007g

(11) Anonymous, “Maxwell 2D-Electromagnetic-field simulation for high performance electromechanical design,” http://www.ansoft.com/products/em/max2d/, March 2007 h

(12) Anonymous, “Image J- Image processing and analysis in JAVA,”

http://rsb.info.nih.gov/ij/, March 2007 i

(13) Auroux P., Iossifidis D., Reyes D., Manz A., “Micro Total Analysis Systems.2. Analytical Standard Operations and Applications,” Analytical Chemistry, 2002; 74: p 2637

107

(14) Bissell M., Sanfilipo F., “Empowering patients with point-of-care testing,” Trends in Biotechnology, 2002; 20 (6): p 269

(15) Beebe D., Mensing G., Walker G., “Physics and applications of microfluidics in

biology,” Annual review of Biomedical engineering, 2002; 4 :p 261 (16) Bassous E., Taub H., Kuhn L., “Ink jet printing nozzle arrays etched in silicon,”

Applied physics letters, 1997; 31: p 135

(17) Burtiz C., Ashwood E., editors, “Tietz Textbook of Clinical Chemistry,” W. B. Saunders Company, 3 edition, 1999.

(18) Bernard A., Michel B., Delamarche E., “Micromosaic immunoassays,” Analytical

Chemistry, 2001; 73: p 8

(19) Bruges J., Roche N., Espinosa J., Rius M., Sastre F., “Evaluation of Ciba Corning ACS:180TM Automated immunoassay system,” Clinical Chemistry, 1995; 40: p 407

(20) Bruin G., “Recent developments in electrokinetically driven analysis on micro

fabricated devices,” Electrophoresis, 2000; 21: p 3931 (21) Bohm S., Olthius W., Bergveld P., “A micromachined double lumen microdialysis

probe connector with incorporated sensor for online sampling,” Sensors and actuators

B, 2000; 63: p 201 (22) Beni G., Hackwood S., “Electrowetting displays,” Applied Physics letters, 1981; 38:

p 207

(23) Chapman T., “Automation on the move,” Nature, 2003; 421: p 661 (24) Chovan T., Guttman A., “Microfabricated devices in biotechnology and

biochemical processing,” Trends in Biotechnology, 2002; 20: p 116 (25) Chan D., “Immunoassay automation- An updated guide to systems,” Academic Press,

SanDeigo, 1996 (26) Cho S., Moon H., Kim C., “Creating, transporting, cutting and merging liquid

droplets by electrowetting-based actuation for digital microfluidic circuits,” Journal of

microelectromechanical systems, 2003; 12: p 70 (27) Chaudhury M., Whitesides G., “How to make water run uphill,” Science, 256: p 1539,

1992 (28) Dempsey E., Diamond D., Smyth M., Urban G., Jobst G., Moser I., Verpoorte E.,

Manz A., Widmer H., Rabenstein K., Freaney R., “Design and development of a

108

miniaturized total chemical analysis system for online lactate and glucose monitoring in biological samples,” Analytical Chimica Acta, 1997; 346: p 341

(29) deMello A., Beard N., “Dealing with real samples: sample pre-treatment in

microfluidic systems,” Lab on a chip, 2003; 3: p 11-N

(30) Delanghe J., Chapelle J., el Allaf M., De Buyzere M., “Quantitative turbidimetric assay for determining myoglobin evaluated,” Ann. Clin Biochem, 1991; 28: p 474

(31) Doherty E., Meagher R., Albarghouthi M., Barron A., “Microchannel wall coatings

for protein separations by capillary and chip electrophoresis,” Electrophoresis, 2003; 24: p 34

(32) Duffy D., Gillis H., Lin J., Sheppard N., Kellogg G., “Microfabricated centrifugal

microfluidic systems: characterization and multiple enzymatic assasy,” Analytical

Chemistry, 1999; 71: p 4669 (33) Dussan E., “On the spreading of liquids on solid surfaces: static and dynamic contact

lines,” Ann. Rev. Fluid Mech., 1979; 11: p 371, (34) Erickson D., Li D., “Integrated microfluidic devices,” Analytica Chimica Acta, 2004;

507: p 11 (35) Eteshola E., Leckband D., “Development and characterization of an ELISA assay in

PDMS microfluidic channels,” Sensors and Actuators B, 2001; 72: p 129

(36) Frost and Sullivan market research report ,“Strategic analysis of point of care testing markets worldwide,” 2003

(37) Gwynne P., Heebner G., “Lab automation and robotics: the brave new world of

24/7 research,”http://www.sciencemag.org/feature/e-market/benchtop/robotfinal.shl, 2002

(38) Gosling J.P., “Immunoassays,” Oxford University Press, UK, 2000 (39) Gijs M., “Magnetic bead handling on chip: new opportunities for analytical

applications,” Microfluidics Nanofluidics, 2004; 1: p 22 (40) Huang Y., Mather E.L., Bell J.L., Madou M., “MEMS based sample preparation

for molecular diagnostics,” Analytical and Bioanalytical chemistry, 2002; 372: p 49 (41) Huels C., Muellner S., Meyer H., Cahill D., “The impact of protein biochips and

microarrays on the drug development process,” Drug Discovery Today, 2002; 7: S 119

109

(42) Hayashi K., Kurita R., Horiuchi T., Niwa O., “Selective detection of l-glutamate using a microfluidic device integrated with a enzyme-modified pre-reactor and an electrochemical detector,” Biosensors and Bioelectronics, 2003; 18: p 1249

(43) Hirano T., “ Biological and clinical aspects of interleukin-6,” Immunology Today,

1992; 11: p 443 (44) Hirano T., “Interleukin-6 and its relation to inflammation and disease,” Clinical

Immunology and Immunopathology, 1990; 62: p S 60 (45) Hirano T., Kishimoto T., “Interleukin-6: Peptide growth factors and their receptors I,”

Springer-Verlag, NewYork, 1990; p 633-665 (46) Jakeway S., de Mello A., Russell E., “ Miniaturized total analysis systems for

biological analysis,” Fresenius’ Journal of Analytical chemistry, 2000; 366: p 525 (47) Kishimoto T., “Interleukin-6 and its receptor: a paradigm for cytokines,” Science,

1992; 258: p 593 (48) Lion N., Reymond F., Girault H., Rossier J., “Why the move to microfluidics for

protein analysis?,” Current opinion in Biotechnology, 2004; 15: p 31 (49) Lee K., “Proteomis: a technology-driven and technology-limited discovery science,”

Trends in Biotechnology, 2001; 19 (6): p 217

(50) Leach A., Wheeler A., Zare R., “Flow injection analysis in a microfluidic format,” Analytical Chemistry, 2003; 75: p 967

(51) Mouradian S., “Lab-on-a-chip: applications in proteomics,” Current opinion in

Chemical Biology, 2002; 6: p 51

(52) Morse M., Clay T., Lyerly H., editors, “Handbook of Cancer Vaccines (Cancer Drug Discovery and Development,” Humana Press; Bk&CD-Rom edition, 2004

(53) Manz A., Graber N., Widmer H.M., “Miniaturized total chemical analysis systems: A novel concept for chemical sensing,” Sensors and actuators B, 1990; 1: p 244

(54) Moser I., Jobst G., Svasek P., Varaharam M., Urban G., “Rapid liver enzyme assay

with miniaturized liquid handling system comprising thin film biosensor array,” Sensors and Actuators B, 1997; 44: p 377

(55) Honda N., Lindberg U., Andersson P., Hoffmann S., Takei H., “Simultaneous

Multiple Immunoassays in a Compact Disc–Shaped Microfluidic Device Based on Centrifugal Force,” Clinical Chemistry, 2005; 51: 1955

110

(56) Nakamura H., Murakami, Yokoyama K., Tamiya E., Karube I., “A compactly integrated flow cell with a chemiluminescent detection FIA system for determining lactate concentration in serum,” Analytical chemistry, 2001; 73: p 373

(57) Nishimoto N., Kishimoto T., “Interleukin-6: from bench side to bedside” Nature

Clinical Practice, 2006; 2: p 619 (58) Pollack M., “Electrowetting based microactuation of droplets for digital

microfluidics,” PhD Thesis, Duke University, Durham, NC, USA, 2001 (59) Pollack M., Shendorov A., Fair R., “Electrowetting-based actuation of droplets for

integrated microfluidics,” Lab on a chip, 2001; 2 (60) Paik P., “Rapid droplet mixers for digital microfluidic systems,” Master’s thesis,

Duke University, 2003a (61) Paik P., Pamula V., Pollack M., Fair R., “Electrowetting-based droplet mixers for

microfluidic systems,” Lab on a chip, 2003b; 3: p 28

(62) Paik P., Pamula V., Pollack M., Fair R., “Rapid droplet mixers for digital

microfluidic systems,” Lab on a chip, 2003c; 3: p 253 (63) Paik P., “Adaptive hot-spot cooling of integrated circuits using digital microfluidics,”

PhD thesis, Duke University, Durham, NC, USA, 2006 (64) Patterson W., Werness P., Payne W., Matsson P., Leflar C., Melander T., Quast S.,

Stejskal J., Carison A., Macera M., Schubert W., “Random and Continuous-Access Immunoassays with Chemiluminiscent detection by Access® Automated analyzer,” Clinical Chemistry, 1994; 40: p 2042

(65) Probstein R., “Physicochemical Hydrodynamics,” Butterworth-Heineman, Boston,

MA, 1989 (66) Piras L., Reho S., “Colloidal gold based electrochemical immunoassays for the

diagnosis of acute myocardial infarction,” Sensors and actuators B, 2005; 111-112: p 450

(67) Petersen K., “Fabrication of an integrated, planar silicon ink-jet structure,” IEEE

Transactions on electron devices,1979; ED-26: p 1918

(68) Qiaojia H., Yuchai T., Weiping L., Shuxuan Y., Xiaopeng L., Bing X., Yushui W., Li L., Zhongyong Z., “Qualitative bed side assay of increased human serum myoglobin by sandwich dot-immunogold filtration for the diagnosis of acute myocardial infarction,” Clinical Chimica Acta, 1998; 273: p 119

111

(69) Ren H., Srinivasan V., Fair R., “Design and testing of an interpolating mixing architecture for electrowetting based droplet on chip chemical dilution,” In Transducers

conference 2003, Boston, MA, June 8-12 2003. (70) Ren H., “Electrowetting based droplet formation,” PhD thesis, Duke University,

Durham, NC, USA, 2004 (71) Reyes D., Iossifidis D., Auroux P., Manz A., “Micro Total Analysis Systems.1.

Introdcution, Theory, and Technology,” Analytical Chemistry, 2002; 74: p 2623 (72) Stephans E.J., “Developing open standards for point of care connectivity,” IVD Tech.,

1999; 5, p 22 (73) Streckfus C., Bigler L., “ Saliva as a diagnostic fluid,” Oral diseases,2002; 8: p 69 (74) Smith J., Osikowicz G., “Abbott AxSYMTM random and continuous access

immunoassay system for improved workflow in the clinical laboratory,” Clinical

Chemistry, 1993; 39: p 2063 (75) Srinivasan V., “A digital microfluidic lab-on-a-chip for clinical diagnostic

applications,” PhD thesis, Duke University, Durham, NC, USA, 2005

(76) Srinivasan V., Pamula V., Fair R., “An integrated digital microfluidic lab-on-a-chip for clinical diagnostics on human physiological fluids,” Lab on a Chip, 2004; 4: p 310

(77) Srinivasan V., Pamula V., Paik P., Fair R., “Protein stamping for MALDI mass

spectrometry using an electrowetting-based microfluidic platform,” Lab on a chip: Platforms, devices and applications, Conf. 5591, SPIE Optics East, Philadelphia, Oct 25-28, 2004

(78) Technote#101, “Proactive microspheres,” Bangs Laboratories, 1999

(79) Technote#301, “Proactive microspheres,” Bangs Laboratories, 1999

(80) Terry S., Jerman J., Angell J., “A gas chromatographic air analyzer fabricated on a

silicon wafer,” IEEE Transactions on Electron Devices, ED-26: p 1880 (81) Taylor R., James T., “Enzymatic determination of sodium and chloride in sweat,”

Clinical Biochemistry, 1996; 29(1): p33

(82) Tudos A., Besselink G., Schasfoort R., “Trends in miniaturized total analysis systems for point of care testing in clinical chemistry,” Lab on a Chip, 2001; 1: p 83

112

(83) Verpoorte E., “Microfluidic chips for clinical and forensic analysis,” Electrophoresis, 2002; 23: p 677

(84) Verpoorte E., “Beads and chips: new recipes for analysis,” Lab on a chip, 2003; 3: p

60N (85) Vo-Dinh T., Cullum B., “Biosensors and biochips: advances in biological and

medical diagnostics,” Fresenius Journal of Analytical chemistry, 2000;366: p 540

(86) Weigl B., Bardell R., Cabrera C., “Lab-on-a-chip for drug development,” Advanced

Drug Delivery reviews, 55: p 349 (87) Wang J., “Electrochemical detection for microscale analytical systems: a review,” Talanta, 2002; 56: p 223 (88) Yoon J., Garrell R., “Preventing bimolecular adsorption in electrowetting based

biofluidic chips,” Analytical chemistry, 2003,75: p 5097 (89) Wright P., “Measurement of insulin secretion. A review of current methods,”

Diabetes, 1968; 17: p 641

113

BIOGRAPHICAL SKETCH

Ramakrishna Sista

Ramakrishna Sista received his Bachelors degree in Chemical Engineering from Andhra

University, India in 2003. Upon graduation, he joined the Chemical Engineering program

at Florida State University to seek a doctoral degree.

Date of Birth - 8th June, 1982

Place of Birth - Visakhapatnam, India