Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Concentration of bacterial pathogens using microfluidic dielectrophoresis systems Lin, Tao 2010

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2010_spring_lin_tao.pdf [ 3.73MB ]
Metadata
JSON: 24-1.0058905.json
JSON-LD: 24-1.0058905-ld.json
RDF/XML (Pretty): 24-1.0058905-rdf.xml
RDF/JSON: 24-1.0058905-rdf.json
Turtle: 24-1.0058905-turtle.txt
N-Triples: 24-1.0058905-rdf-ntriples.txt
Original Record: 24-1.0058905-source.json
Full Text
24-1.0058905-fulltext.txt
Citation
24-1.0058905.ris

Full Text

Concentration of Bacterial Pathogens Using Microfluidic Dielectrophoresis Systems by  Tao Lin Bachelor of Marine Science, Guangdong Ocean University, 2007  A THESIS SUBMITTED IN PARTIAL FULFILLENT OF THE REQUIREMENT FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE in The Faculty of Graduate Studies (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2010 © Tao Lin, 2010  ABSTRACT Over the course of human history, people have been always troubled by unpredictable outbreaks of pathogenic bacteria. A rapid, specific and durable diagnostic sensor for detecting pathogens in a timely fashion is essential for identifying unknown infectious diseases and reacting by offering proper patient management and public health involvement without delay. Development of a microfluidic system incorporating interdigitated electrodes for characterizing concentration, purification and differentiation of pathogenic bacteria is presented. Polystyrene microspheres, Salmonella typhimurium, Campylobacter jejuni, Escherichia coli and Mycobacterium smegmatis are trapped by the positive dielectrophoretic force produced by the microelectrodes deposited and arrayed on a glass substrate covered with a PDMS microchannel. Diluted buffer used for producing positive dielectrophoresis for capturing cells is selected after evaluating the viability of the cells suspended in different buffers. Trapping efficiencies of Escherichia coli and Mycobacterium smegmatis are acquired by comparing input concentration and output concentration from plating results. Effective conductivities of both Escherichia coli and Mycobacterium smegmatis are calculated by obtaining optical density measurements corresponding to certain buffers with a range of different conductivity values used for trapping. Separation of a mixture with Escherichia coli and Mycobacterium smegmatis is  ii  achieved through the microconcentrator, although the results conflict with the calculated cellular conductivity values. The microfluidic DEP concentrator is shown to be an effective tool for studying and separating bacterial populations, and the results clearly indicate exciting directions for future work.  iii  TABLE OF CONTENTS ABSTRACT ....................................................................................................................... ii TABLE OF CONTENTS ................................................................................................ iv LIST OF TABLES ........................................................................................................... ix LIST OF FIGURES .......................................................................................................... x ACKNOWLEDGEMENTS ........................................................................................ xviii DEDICATION................................................................................................................. xx Chapter 1: Introduction ................................................................................................... 1 1.1 Research Background ............................................................................................... 1 1.2 Conventional Pathogenic Assays .............................................................................. 4 1.2.1 Enrichment Culture ............................................................................................ 4 1.2.2 Immunological Detection Methods.................................................................... 5 1.2.3 Molecular Biology Detection Methods .............................................................. 6 1.3 Microfluidic Biodiagnosis ........................................................................................ 7 1.4 Research Motivations................................................................................................ 9 Chapter 2: Applications of Dielectrophoresis .............................................................. 11 2.1 Introduction of Dielectrophoresis ........................................................................... 11  iv  2.2 Dielectric Properties of Biological Particles ........................................................... 18 2.3 Advantages of Applying Microdielectrophoresis ................................................... 25 2.3.1 Previous Works Overview ............................................................................... 26 2.3.2 Advantages and Limitations of Positive DEP .................................................. 27 2.3.3 Electrode Design .............................................................................................. 29 2.3.4 Summary .......................................................................................................... 30 Chapter 3: Microfabrication .......................................................................................... 32 3.1 Techniques and Facilities ........................................................................................ 32 3.1.1 Photolithography .............................................................................................. 33 3.1.2 Electron-beam (e-beam) Evaporation .............................................................. 35 3.1.3 Plasma-enhanced Chemical Vapor Deposition (PECVD) ............................... 37 3.2 Polydimethylsiloxane (PDMS) Channel Chip ........................................................ 39 3.2.1 Mask Design .................................................................................................... 39 3.2.2 Fabrication ....................................................................................................... 41 3.3 Interdigitated Microelectrodes Chip ....................................................................... 44 3.3.1 Mask Design .................................................................................................... 44 3.3.2 Fabrication ....................................................................................................... 48 3.4 Device Packaging.................................................................................................... 51 3.4.1 PDMS-glass Bonding....................................................................................... 51  v  3.4.2 Chip Assembly ................................................................................................. 54 Chapter 4: Materials and Methods ............................................................................... 56 4.1 Experimental Materials ........................................................................................... 56 4.1.1 Suspending Fluid Media .................................................................................. 56 4.1.2 Polystyrene Microsphere ................................................................................. 59 4.1.3 Salmonella typhimurium .................................................................................. 60 4.1.4 Campylobacter jejuni ....................................................................................... 62 4.1.5 Escherichia coli ............................................................................................... 64 4.1.6 Mycobacterium smegmatis............................................................................... 66 4.2 Testing Particles Preparation and Storage .............................................................. 68 4.2.1 Non-biological Particles................................................................................... 68 4.2.2 Biological Particles .......................................................................................... 68 4.3 Measurement Apparatuses ...................................................................................... 69 4.3.1 Conductivity Meter .......................................................................................... 69 4.3.2 UV Spectrophotometer .................................................................................... 70 4.3.3 Fluorescence Microscope................................................................................. 71 4.4 Experimental Setup ................................................................................................. 72 4.5 Experimental Procedures ........................................................................................ 74 4.5.1 Concentration Efficiency Evaluation ............................................................... 74  vi  4.5.2 Particle Effective Conductivity Assessment .................................................... 76 4.5.3 Mixed Sample Separation ................................................................................ 77 Chapter 5: Results and Discussion ................................................................................ 78 5.1 Standard Curves ...................................................................................................... 78 5.1.1 Salmonella typhimurium .................................................................................. 78 5.1.2 Escherichia coli ............................................................................................... 81 5.1.3 Campylobacter jejuni ....................................................................................... 83 5.1.4 Mycobacterium smegmatis............................................................................... 85 5.2 Viability of Bacterial Cells in Diluted Buffer ......................................................... 87 5.2.1 Salmonella typhimurium in Diluted PBS and HEPES ..................................... 88 5.2.2 Escherichia coli in Diluted PBS and HEPES .................................................. 91 5.2.3 Campylobacter jejuni in Diluted PBS ............................................................. 93 5.2.4 Mycobacterium smegmatis in Diluted HEPES ................................................ 94 5.3 DEP Capturing in Diluted PBS ............................................................................... 95 5.3.1 Microspheres Capturing ................................................................................... 95 5.3.2 Salmonella typhimurium Capturing ................................................................. 97 5.3.3 Campylobacter jejuni Capturing ...................................................................... 98 5.4 Capturing Efficiency Evaluation........................................................................... 100 5.4.1 Escherichia coli Capturing ............................................................................ 100  vii  5.4.2 Mycobacterium smegmatis Capturing............................................................ 102 5.5 Effective Conductivities Estimation ..................................................................... 104 5.6 Mixed Sample Partitioning ................................................................................... 109 5.7 DEP Capturing in Bodily Fluid............................................................................. 116 Chapter 6: Conclusions ................................................................................................ 120 6.1 Summary ............................................................................................................... 120 6.2 Future Work .......................................................................................................... 122 BIBLIOGRAPHY ......................................................................................................... 125 APPENDICES ............................................................................................................... 135 Appendix A. Schematic of SMT Board ...................................................................... 135 Appendix B. Raw Data of Growth Curves and Calibration Curves ........................... 136 Appendix B. Raw Data of Viability Assessment ........................................................ 140 Appendix C. Raw Data of Trapping Efficiency Evaluation ....................................... 143 Appendix D. Raw Data of Effective Conductivities Estimation ................................ 145 Appendix E. UBC Research Ethics Board's Certificates of Approval ....................... 147  viii  LIST OF TABLES Table 2-1: ―Comparison of advantages and disadvantages of p-DEP and n-DEP approaches to trapping cells.‖ [22] ...................................................................................................... 28 Table 4-1: Spectra of fluorescent particles and microscope filters. .................................. 72  ix  LIST OF FIGURES Figure 2-1: Difference between electrophoresis and dielectrophoresis. ........................... 12 Figure 2-2: Two modes of dielectrophoretic forces. A. Positive DEP; B. Negative DEP.16 Figure 2-3 Force balance of a homogeneous sphere subjected to laminar flow channel with interdigitated electrodes array. .......................................................................................... 17 Figure 2-4:Dielectric shell model (Rm: resistance of suspending medium; R1 and R2: resistances of cell wall and membrane; Z: impedance of cytoplasm; Cm: capacitance of suspending medium; C1 and C2: capacitances of cell wall and membrane). .................... 20 Figure 2-5: ―Simulated dielectrophoretic properties for: (a) point particles; (b) solid particles; (c) particles with a single compartment surrounded by a thin envelope; and (d) particles with two concentric compartments surrounded by thin envelopes. εs and σs are per ‖ [28] .................................................................................................................................... 22 Figure 2-6: Cell wall structures of Gram-positive and Gram-negative cells (Source: Dr. Kaiser's Microbiology Home Page). ................................................................................. 24 Figure 3-1: Cleanroom in AMPEL at UBC. (Source: AMPEL website.) ........................ 33 Figure 3-2: Basic photolithography process flow chart. A. Photoresist coating; B. UV exposing; C. Developing; D. Etching. .............................................................................. 34  x  Figure 3-3: Canon Mask Aligner (Model: PLA-501F). (Source: AMPEL website.) ....... 35 Figure 3-4: Electron-beam evaporation schematic. .......................................................... 36 Figure 3-5: E-beam evaporation system. (Source: AMPEL website.).............................. 37 Figure 3-6: PECVD schematic.......................................................................................... 38 Figure 3-7: PECVD/RIE: Trion. (Source: AMPEL website.) .......................................... 39 Figure 3-8: Masks of PDMS channel. A. Dual channel design; B. Single channel design. ........................................................................................................................................... 41 Figure 3-9: Costumed-design SMT board and strip connector soldered on acrylic clamp. ........................................................................................................................................... 45 Figure 3-10: Masks of electrodes. A. Juxtaposed sets of electrodes for dual channel design; B. Tandem sets of electrodes for single channel design. .................................................. 47 Figure 3-11: Ozone reaction mechanism. ......................................................................... 52 Figure 3-12: PDMS surface chemical structure transformation. ...................................... 53 Figure 3-13: Fully assembled DEP chips with electrode substrates bonded to PDMS channels............................................................................................................................. 54 Figure 3-14: Assembled and packaged DEP chip located within the SMT board. Fluidic tubing and BNC connectors are visible. ........................................................................... 55  xi  Figure 4-1: Chemical structure of HEPES. ....................................................................... 58 Figure 4-2: Electron microscope image of S. typhimurium. (Source: Volker Brinkmann, Max Planck Institute for Infection Biology, Berlin, Germany.) ....................................... 61 Figure  4-3:  Electron  microscope  image  of  Campylobacter  jejuni.  (Source:  Virginia-Maryland Regional College of Veterinary Medicine, Blacksburg, Virginia.) ... 63 Figure 4-4: Electron microscope image of E. coli. (Source: Rocky Mountain Laboratories, NIAID, NIH.) .................................................................................................................... 65 Figure 4-5: Electron microscope image of M. smegmatis. (Source: Molecular Microbiology, Victoria University of Wellington, New Zealand.) .................................. 66 Figure 4-6: Experimental setup. ........................................................................................ 73 Figure 5-1: Sample growth curve of S. typhmurium growing in LB broth with tetracycline for 200 minutes. ................................................................................................................ 79 Figure 5-2: Sample calibration curve of S. typhmurium growing in LB broth with tetracycline. ....................................................................................................................... 80 Figure 5-3: Sample growth curve of E. coli growing in LB broth with ampicillin for 200 minutes. ............................................................................................................................. 81 Figure 5-4: Calibration curve of E. coli growing in LB broth with ampicillin. (Error bars were presented but too small to perceive.)........................................................................ 82  xii  Figure 5-5: Calibration curve of C. jejuni growing in BHI plus 5% NBS broth with vancomycin and trimethoprim. (Error bars were presented but too small to perceive). ... 84 Figure 5-6: Calibration curve of C jejuni growing in Mueller-Hinton broth with vancomycin and trimethoprim. (Error bars were presented but too small to perceive). ... 85 Figure 5-7: Calibration curve of M. smegmatis growing in Middlebrook 7H9 Broth. (Error bars were presented but too small to perceive). ................................................................ 86 Figure 5-8: Viability of S. typhimurium in 0.2 mM PBS plus 10% dextrose. .................. 89 Figure 5-9: Viability of S. typhimurium in 0.4 mM HEPES plus 10% dextrose. ............. 90 Figure 5-10: Viability of E. coli in 0.2 mM PBS plus 10% dextrose. .............................. 91 Figure 5-11: Viability of E. coli in 0.4 mM HEPES plus 10% dextrose. ......................... 92 Figure 5-12: Viability of C. jejuni in 0.2 mM PBS plus 10% dextrose. ........................... 93 Figure 5-13: Viability of M. smegmatis in 0.2 mM HEPES plus 10% dextrose. ............. 94 Figure 5-14: Microspheres trapping in 0.2 mM PBS plus 10% dextrose. A. Trapping under 20 Vp-p and 1kHz with a flow rate of 100 μL/hr; B. 20 Vp-p and 1kHz with a flow rate of 500 μL/hr; 20 Vp-p and 3kHz with a flow rate of 100 μL/hr; 20 Vp-p and 3kHz with a flow rate of 500 μL/hr. ......................................................................................................................... 96  xiii  Figure 5-15: S. typhimurium trapping in 0.17 mM PBS plus 10% dextrose. A. Trapping under 5 Vp-p and 3 MHz with a flow rate of 100 μL/hr for 3 minutes; B. Trapping under 10 Vp-p and 3 MHz with a flow rate of 100 μL/hr for 3 minutes. .......................................... 98 Figure 5-16: C. jejuni trapping in diluted PBS with different conductivities plus 10% dextrose. A. Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.17 mM buffer; B. Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.2 mM buffer; Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.28 mM buffer. .......................................... 99 Figure 5-17: E. coli trapping efficiency in diluted HEPES of 40 μS/cm and 120 μS/cm plus 10% dextrose with a flow rate of 150 μL/hr. .................................................................. 101 Figure 5-18: M. smegmatis trapping efficiency in diluted HEPES of 40 μS/cm and 120 μS/cm plus 10% dextrose with a flow rate of 150 μL/hr. ............................................... 103 Figure 5-19: Values of absorbance change of E. coli as a function of the conductivities’ values of suspending media. ........................................................................................... 105 Figure 5-20: Values of absorbance change of M. smegmatis as a function of the conductivity values of the suspending media. ................................................................ 106 Figure 5-21: Values of the suspending media conductivities as a function of (k - ΔA) / (2ΔA+k) for E. coli. ........................................................................................................ 107  xiv  Figure 5-22: Values of the suspending media conductivities as a function of (k - ΔA) / (2ΔA+k) for M. smegmatis. ............................................................................................. 108 Figure 5-23: E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose individually. A. E. coli trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B. ......................................................................................................................................... 111 Figure 5-24: E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose individually. A. E. coli trapped on the second set of electrodes at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the second set of electrodes at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B. ....................................................................................................................... 112 Figure 5-25: Mixture of E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose together. A. E. coli trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μl/hr; B. M. smegmatis trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B. ..................................................................................................................................... 114 Figure 5-26: Mixture of E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose together. A. E. coli trapped on the second set of electrode at 8 Vp-p and  xv  900 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the second set of electrode at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B. ......................................................................................................... 115 Figure 5-27: E. coli trapping in diluted CSF at 120 μS/cm. A. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 5X magnification; B. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 5X magnification; C. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 5X magnification; D. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150μL/hr for 300 seconds with 10X magnification; E. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 10X magnification; F. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 10X magnification. ......................................................................................................................................... 117 Figure 5-28: M. smegmatis trapping in diluted CSF of 120 μS/cm. A. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 5X magnification; B. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 5X magnification; C. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 5X magnification; D. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 10X magnification; E. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 10X magnification; F. Trapping under 15  xvi  Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 10X magnification. ......................................................................................................................................... 118 Figure 6-1: ―Device overview. (A) Illustration of the assembled devices, showing the PDMS channel and gold IDEs on the Pyrex substrate. (B) Illustration of particle trapping in laminar flow and chaotic flow. Without the mixer, only a portion of particles carried by the laminar flow are exposed to DEP and trapped by electrodes. Chaotic flow generated by a passive micromixer circulates the particles and exposes more particles to the DEP region. (C) Schematic of the three kinds of micromixer geometries described in this work. From left to right: slanted groove micromixer (SGM), herringbone micromixer (HM), and staggered herringbone micromixer (SHM).‖ [65] .......................................................... 123  xvii  ACKNOWLEDGEMENTS I would like to thank many people without whom this project and I would not have evolved to the state where we are. First and foremost, I would like to express my great appreciation to my supervisor, Dr. Eric Lagally, for his guidance, assistance, patience, inspiration and edificatory support I have been fortunate enough to receive. I owe special thanks to Dr. Patrick Tang from BC Centre for Disease Control and Dr. Brett Finlay from Michael Smith Laboratory at UBC for always having time, knowledge and patience to offer technical support and experimental materials to this work; and to Dr. Karen Cheung from Biomedical Engineering at UBC for sharing the lab equipments and being generous with her time and knowledge; and Arash Takshi for designing the crucial device and assist in answering any technical question; and the postdoctoral fellows, Matthew Croxen and Hongbing Yu from Finlay lab, for being generous with their intellectual and tutorial assistances; and Jonas Flueckiger from Cheung lab for helping and teaching me the microfabrication in cleanroom. Additional thanks to my colleagues of Lagally lab — Tony Yang, Jie Yu, Roza Bidshahri, Bernard Coquinco, Eric Ouellet, Jose Villegas and Louise Lund for being supportive in my project. For project funding, thanks to Canada Foundation for Innovation and Canadian lung association as well as BC Centre for Disease Control.  xviii  The last but not the least, I would like to appreciate my family, especially my Mom LEUNG Laiping , for her everlasting encouragements for me.  xix  DEDICATION  To My Parents & Aunt Yukching Lam  xx  CHAPTER 1: INTRODUCTION 1.1 Research Background As scientists acquire more in-depth knowledge about microorganisms, a extensive diversity of bacteria have been shown to be valuable and inoffensive to human beings[1]. Nevertheless, there are also many pathogenic strains of bacteria, which can cause contagious diseases and may even lead to death in some serious cases. Pathogenic bacteria are present almost in every area where they are able to be a major cause of significant morbidity and mortality, including food-borne illness, soft tissues infections, biological warfare, and lung disease. Food-related outbreaks happen globally both in developing and developed countries due to the toxins produced by the pathogenic intestinal microorganisms, including Salmonella, Escherichia, Campylobacter, Clostridium. The infectious sources of Salmonella can be contaminated water or food, especially meat, poultries, raw milk, eggs, vegetables, fruit and processed foods. Salmonella serotype Typhimurium and Enteritidis, Newport and Heidelberg are the most prevalent in western industrial countries accounting for about 60% of the infection cases reported to the Centre for Disease Control [2]. Escherichia, especially the shiga toxin (stx)-producing strains of Escherichia coli (STEC) including most the predominant serotypes O157:H7, O157:NM, and the non-O157  1  serotypes O26:H11, O111:NM, O103:H2, and O145:NM, are important food-borne pathogens leading to watery diarrhea and hemorrhagic colitis followed with hemolytic-uremic syndrome (HUS) [3]. Particularly, children younger than five are the highest susceptible populations of HUS cases caused by STEC O157. However, as patient age exceeds 60, the mortality rate increases regardless of whether HUS is a serious complication [4]. Further, the pathogenic strain of Campylobacter including Campylobacter jejuni are among the most significant cause of food-borne gastroenteritinal disease in the world. Every year, about 400 to 500 million diarrhea cases worldwide are triggered by Campylobacter jejuni. Although mediated gastroenteritinal infections caused by Campylobacter jejuni are self-limiting, certain ones has been reported to connect to Guillain-Barre´Syndrome (GBS), a severe disease of peripheral nervous system [5]. Another etiological bacterium, namely Clostridium difficile, has been capturing more and more attention because they attack susceptible patients, especially the ones receiving antibiotic treatment in hospital, and cause them antibiotic-associated diarrhea (AAD), as well as pseudomembranous colitis (PMC), in serious cases resulting in death. Increasing incidences of nosocomial infections as well as community-associated infections are recognized and documented in North America and Europe due to the recent discovery of the hypertoxigenic strain [6, 7]. In Canada, Clostridium difficile infection (CDI) had a higher rate in 2009 than in both 2006 and 2007. 16 people died in the outbreak of  2  Clostridium difficile in Quebec hospitals in 2006 and 460 patients have died because of the CDI in Ontario [6]. Soft  tissues  infections  caused  by  Staphylococcus  aureus  are  both  community-associated and healthcare-associated. In particular, methicillin-resistant Staphylococcus aureus (MRSA) is increasingly epidemic leading to more significant morbidities and mortalities [7]. In terms of biological warfare, agents can be developed from pathogens designed as weapons to attack the organisms including humans. Because of the invisibility and latency of biowarfare, infectious symptoms from victims caused by certain pathogens require a few days to present. The most frequent agent applied in bioterrorist attacks is anthrax derived from Bacillus anthracis [8]. Perhaps the most significant bacterial pathogens are those causing lung diseases, namely tuberculosis (TB), among which Mycobacterium tuberculosis is the most common causative agent. From a recent report from the World Health Organization (WHO), 9.4 million new TB cases appeared and 1.8 million people died due to TB in 2008 alone. TB remains a top killer of human beings, especially patients infected with human immunodeficiency virus (HIV). In addition, more and more infection incidences are associated with multidrug-resistant TB (MDR-TB) or even extensively drug-resistant TB (XDR-TB), which will dramatically increase the difficulty of medical treatment and disease control [9].  3  Therefore, effective analytical methods that operate in a timely fashion are indeed critical to reduce the time needed for initial pathogenic detection before desired patient management and appropriate healthcare intervention may be obtained. Current methods are unable to achieve this goal; however, a novel detection method coupling concentration, separation and combination with more specific molecular detection down to the strain level is both possible and definitely preferred.  1.2 Conventional Pathogenic Assays Normally, there are several traditional methods for detecting bacterial pathogens, including enrichment culture, immunological method such as latex agglutination test (LAT), and molecular biology detection methods. These methods still occupy positions in the pathogenic detection field, even the time-consuming and complicated procedures. However, individual use of each of these methods is insufficient, and a combination of techniques usually provides a more robust detection scheme.  1.2.1 Enrichment Culture Liquid or solid media culture has been used to proliferate particular bacteria for decades. At present, specimens from blood, urine or cerebrospinal fluid are cultured on specific media for a few days. This allows access to quantitative inspection of the sample  4  by colony counting with the purpose of measuring if the bacteria exceed the minimum infectious dose. Treatments will then be applied depending on the counting results. This process takes a long time. For example, a negative result of Campylobacter infection requires waiting between 4 and 9 days and a positive one requires 14 to 16 days for confirmation [10]. Additionally, it is hard visualize bacterial clusters growing on plates when they have similar or confused morphologies that affect the identifications. Therefore, various media with selective and specific features are developed for particular species of bacteria. Commonly, specific compositions are added to particularly colorize the growing targeted bacteria [11], degrade the untargeted ones or maintain the growth of selected ones while inhibiting the others from further growing. This will help to easily differentiate certain species of interest by visible eyes. However, some pathogens, such as Mycobacterium tuberculosis, only need very low concentrations to cause disease and enrichment culture hinders effective detection due to survivability losses during sample transportation.  1.2.2 Immunological Detection Methods Applying the principle of antibody-antigen interactions can be a robust diagnostic technique for detecting various targets. One of the laboratory instances is LAT, which is method using polystyrene pre-beads coated with antibody to check for target antigens  5  through immunological reaction and agglutination. The whole combination will clump and settle due to co-agglutination when an appropriate antibody reacts with multiple target antigens. At this point, visual examination can be performed. Nevertheless, LAT is limited in that has been reported to be an unreliable and insensitive technique for testing bacterial meningitis [12, 13]. Another immunology-related technique with better performance than LAT is the enzyme-linked immunosorbent assay (ELISA) test. In the case of ELISA, detectable signal can be obtained from an easily evaluated enzyme bound to the antibody. And this antibody has been specifically associated with an unknown antigen immobilized on a solid surface [14]. The ELISA has been established as an analytical tool widely exploited in plenty of industries including medicine and plant pathology but still requiring significantly practical time.  1.2.3 Molecular Biology Detection Methods Molecular techniques for pathogenic bacteria detection usually involve a nucleic acid-based assay method, combining polymerase chain reaction (PCR) with gel electrophoresis analysis. A target DNA sequence can be amplified and quantified in vitro through PCR. This can be used to identify the existence of specific genetic material or any anomaly including mutations, deletions and recombinations in targeted bacteria. For  6  example, PCR can be an analytical tool for the diagnosis and appraisal of genetic disease or tumour (cancer) in medical industries and the identification of bacterial, viral or fungicidal infection. One of the major benefits of PCR is that it also acts as a product line to produce lots of specific target genes since it is able amplify and quantify target DNA sequence exponentially in a time-saving manner. There are several types of PCR possessing different performances that are applied to various fields. For instance, reverse transcriptase PCR is used to overcome the restrictions of distinguishing viable cells from dead cells, which have all DNA. In this case, the single-stranded DNA in target bacteria can be synthesized from RNA in the direction of 5’ to 3’ by the particular ability of reverse transcriptase following with the specific detection from several genes appear when bacteria are growing [15]. However, the bottleneck of primer design restriction for target DNA sequence has not yet been overcome, and reagents used in PCR are relatively expensive.  1.3 Microfluidic Biodiagnosis The miniaturization and simplification of conventional techniques on a biochip have been rapidly developed as promising tools for pathogenic detection, coupling  7  multidisciplinary fields including biology, chemistry, optics, piezoelectricity, magnetic and micromechanics. The biochip has the potential to achieve high sensitivity and specificity analysis as well as rapid detection. Results from testing multiple target samples can be acquired through one single experiment, which will increase the detection efficiency. Recently, a microfluidic biosensor integrated with (surface plasmon resonance) SPR and fluorescence was developed by Zordan et al. Applying the specificity of an antibody against the antigen of targeted pathogen, the device was shown to selectively detect magnetically preconcentrated pathogenic strains of Escherichia coli O157:H7 by SPR on a functionalized gold slide and fluorescence imaging which shows potential to test food-borne pathogens in microdevice [16]. In 2008, Beyor et al. introduced an integrated microsystem performing capture of immunomagnetic-labeled pathogens from a dilute sample by off-chip PCR and capillary electrophoretic assay. 70% capture efficiency of Escherichia coli at a concentration of 100 colonies formed units per microliter was achieved with a detection limit of 2 colonies formed units per microliter [17]. Another example of detecting pathogens in a microfluidic system was presented by Chang in 2007. A dielectrophoresis based microfluidic platform was fabricated for screening multitarget pathogen with high-throughput. The versatility of this device included filtering, focusing, sorting and trapping multiple targets in a continuous flow[18].  8  In general, microfluidic systems have high potential to be developed as the major detection methods for pathogenic diagnosis since the flexibility and feasibility of designing and creating a uniform microfluidic system can incorporate multiple testing methods integrated in a miniaturized scale.  1.4 Research Motivations The research presented in this thesis is aimed at developing an integrated microsystem with microfabricated electrodes inducing non-uniform electric fields for characterizing and optimizing the concentration and separation of microorganisms of interest. This allows building up a model for the assistance of detecting and diagnosing pathogenic bacteria. Conventionally, depending on the species of bacteria, it takes a relatively long time to collect enough enrichment samples purified from raw clinical specimens before commencing genetic identification and analysis in downstream. This will critically limit and constrain a rapid and proper clinical responds to an unpredictable outbreak of specific pathogen when it happens. Therefore, a simple, quick and sensitive diagnostic device having the functionalities of specific concentration and selective purification is ideally preferred. Dielectrophoresis is one of the promising methods to overcome the insufficiencies of traditional detection methods. Combining the advantages of microfluidic techniques, an integrated  9  microdielectrophoresis concentrator is presented to perform a rapid characterization of multiple species of bacterial pathogens from pure samples. Moreover, the ultimate intention of this microsystem is to construct an interface between raw clinical samples and downstream diagnostic detection in a rapid and specific manner. Subsequently, the abilities of high throughput, low power, low cost as well as portability are desired for extensive use outside of the laboratory setting, especially in scenes of pathogenic outbreak in developing countries where on-site laboratory facilities are absent.  10  CHAPTER 2: APPLICATIONS OF DIELECTROPHORESIS Combining the advances of microfluidic techniques and the properties of dielectrophoresis (DEP), a micro-scale system is developed to provide a concentration of target bio-particles. One critical component in this concentrator is the electrodes, which are employed to produce the electric field and capture bio-particles.  2.1 Introduction of Dielectrophoresis The concept of DEP was initially defined by Dr. Pohl in 1951 [19], which refers to the motion of charge-neutral particles in non-uniform electric fields arising from differences in dielectric properties between target particles and their surrounding fluid medium [20]. As shown in Figure 2-1, in a uniform electric field a neutral particle is merely polarized and remains immobile in uniform electric field, while a charged particle moves along the electric field lines. This phenomenon is simply electrophoresis. However, the neutral particle will be polarized and this will induce a dipole moment the same as or opposite the electric field direction when subjected to a non-uniform electric field, where dielectrophoresis arises. The polarized particles suspended in a fluid medium will then  11  move to the strongest or weakest field regions under the influence of their induced dipole moments and the electric field, resulting in a separation.  Figure 2-1: Difference between electrophoresis and dielectrophoresis.  The force derived from the neutral particles surrounded by a fluid medium under the influence of a non-uniform electric field can be expressed as FDEP. For spherical particles, the time-average DEP force is elucidated as [21, 22],  3  𝐹DEP = 2πr 𝜀𝑚 Re  ∗ 𝜀 𝑝∗ −𝜀 𝑚 ∗ 𝜀 𝑝∗ +2𝜀 𝑚  ∇E  2  (2-1)  where r is the particle radius, εm is the real part of the permittivity of the suspending medium, εp* is the complex permittivity of the particleand εm* is complex permittivity of  12  the surrounding medium, and E is the root mean-square electric field. Here {(εp*- εm*)/ (εp*+2 εm*)} is the Clausius-Mossotti factor, K, which is a measurement of the effective polarizability of the particle.  𝜀∗ = 𝜀 +  𝜎 𝑗𝜔  (2-2)  where ε* is equal to εp* or εm*, ε is the dielectric constant, σ is the electrical conductivity, ω is the field frequency, and j is the imaginary number. A characteristic time constant is necessary to produce polarization charges on the particle, described as,  𝜏=  𝜀 𝑝 +2𝜀 𝑚 𝜎𝑝 +𝜎𝜀 𝑚  (2-3)  Where εp is permittivity of the particle and εm is permittivity of the surrounding medium. σp and σm are the conductivities of the particle and the surrounding medium, respectively. Therefore, τ has to be larger than the frequency of the applied electric field so that the particles will be able to have enough time to reflect the electric field properties. Rearranging Equation 2-3 and substituting into {(εp*- εm*)/ (εp*+2εm*)} yields an expanded expression for K:  13  𝐾=  (𝜎𝑝 −𝜎𝑚 ) (1+𝜔 2 𝜏 2 )( 𝜎𝑝 +2𝜎𝑚 )  +  𝜔𝜏 2 (𝜀 𝑝 −𝜀 𝑚 ) (1+𝜔 2 𝜏 2 )( 𝜀 𝑝 +2𝜀 𝑚 )  (2-4)  where ω = 2πƒ. Depending on the value of K, DEP can thus be divided up into two modes and behaves differently in each mode. When K > 0, the particle moves towards the electrodes and the strongest-electric-field area by positive DEP (p-DEP). The particle is repelled or levitated from the electrode when negative DEP (n-DEP) is dominating with K ˂ 0. When εp* is much larger or less than the εm*, the range of K can range from -0.5 to 1.0 in terms of the expression of K shown in Equation 2-1. According to Equation 2-4, whether K is larger or less than zero depends on (σp-σm) and (εp-εm). Therefore, controlling the values of (σp-σm) and (εp-εm) provides a way of operating in either the p-DEP or n-DEP modes. (σp-σm) becomes the main factor to control the DEP performance when lower frequency is applied. When the applied frequency is intermediate, both (σp-σm) and (εp-εm) are of importance. (εp-εm) will only dominate under the highest frequency condition. Counterion adsorption and neutralization between the particle surface and the surrounding medium occur based on the principle of opposite charge attraction. Therefore, more charges on the surface of the particle with stronger polarizability may be induced and these charges generated from the particle itself will work as the primary control factor, resulting in a dipole moment directed from negative charges to positive charges at the  14  interface between the particle and the medium. A net force orienting the particle to the maximum electric field will then be produced when the direction of dipole moment is the same as the one of electric field, named positive dielectrophoretic force (Figure 2-2A). On the other hand, induced charges from the medium with stronger polarizability than the particle itself may dominate in a non-uniform electric field. The particle will then be repelled toward electric field minima since the direction of dipole moment is opposite to the one of electric field. This is negative dielectrophoretic force (Figure 2-2B).  15  Figure 2-2: Two modes of dielectrophoretic forces. A. Positive DEP; B. Negative DEP.  In this work, p-DEP was applied as the main approach to manipulate the tested particles. Understanding the force balance of a homogeneous particle placed into a flow channel is fundamental in order to modify the DEP force to achieve better trapping performance.  16  Figure 2-3 Force balance of a homogeneous sphere subjected to laminar flow channel with interdigitated electrodes array.  A cross-section of a laminar flow channel incorporated with gold electrodes used to produce p-DEP is shown in Figure 2-3. The uniform sphere is acted on by buoyancy force, DEP force and fluid, where fluid force is the resultant force of drag force and flow force. Due to the influence of the force balance, the composition of the velocity of the particle can be divided up into two components  𝑣𝑥 = 𝑢𝑥 +  𝐹𝐷𝐸𝑃 ,𝑥 6𝜋𝜂𝑟  (2-4)  where,  𝑢𝑥 =  1 𝑝 2𝜂 𝑙  (𝑑2 − 𝑦 2 )  While on the vertical direction (y axis),  17  (2-5)  𝑣𝑦 =  𝐹𝐷𝐸𝑃 ,𝑦 +𝐹𝐵𝑢𝑜𝑦𝑎𝑛𝑐𝑦 6𝜋𝜂 𝑟  (2-6)  2.2 Dielectric Properties of Biological Particles Particles with resistance to electric current flow ordinarily have zero total charge such that uniform electric fields cannot cause motion. However, they may have polarizabilities and produce dipole moments when under the influence of an imposed electric field [23]. Bio-particles, such as biological cells, are thus neutral elements possessing dielectric properties, which contain permanent dipoles themselves and dipoles induced by applied non-uniform electric field. Cells are capable of being polarized because of different electrical properties of materials enclosing in their neighboring structures. The research in 1937 of Curtis and Cole [24] demonstrated that the electrical properties of the cell wall are essential for the dielectric response in applied electric fields. Different types of cells have various structures. In this study, bacterial cells are the principal targets, whose compositions of cell walls are mainly peptidoglycan. It is highly insulating because of its constitution of lipid and many proteins. In addition, many organelles covered with membranes and charge molecules are located inside of the cell, which will contribute 107 times higher conductivity than the cell membrane [25]. The study of the dielectric properties of cells becomes more complicated due to the physiological and  18  morphological variances of different species of cells under numerous growing and living conditions. Therefore, it is beneficial to apply dielectric models on different kinds of cells to study their dielectric properties. Assuming all major structures of a cell can be represented as different resistors and capacitors in a circuit, an equivalent dielectric shell model [26] is illustrated in Figure 2-4.  19  Figure 2-4:Dielectric shell model (Rm: resistance of suspending medium; R1 and R2: resistances of cell wall and membrane; Z: impedance of cytoplasm; Cm: capacitance of suspending medium; C1 and C2: capacitances of cell wall and membrane).  The model is idealized to simulate the cell structure in order to realize and measure experimental processes. Thus each component in the circuit is not required to be exactly equivalent to the physical structures in a cell. Evidence has been demonstrated that the simulated model is legitimately accurate [27] and can be used to develop other shell models for ―point particles, solid particles, particles  20  with a single compartment surrounded by a thin envelope, and particles with two concentric compartments surrounded by thin envelopes‖ [28], as shown in Figure 2-5.  21  Figure 2-5: ―Simulated dielectrophoretic properties for: (a) point particles; (b) solid particles; (c) particles with a single compartment surrounded by a thin envelope; and (d) particles with two concentric compartments surrounded by thin envelopes. εs and σs are per ‖ [28]  22  Each type of particle model has its own structures as well as size, which give variations in the number of parameters (permittivity and conductivity) that need to be considered. A corresponding spectrum of K factor versus frequency can be created showing the unique features of different types of particles representing distinctive dielectric responses in applied electric fields. From Figure 2-5, all the spectra confirm that the K factor is frequency dependent that applied frequency plays a crucial part in discriminating particles with various dimensions and compositions resulting in distinct polarizabilities and dielectrophoretic effects. For example, whether a covered membrane encircles the particle or not can provide a distinguishing signal to identify one from another. In the bacterial world, there are numerous species with different structures and sizes but sharing similar constructions of cytoplasm and cytoplasmic membrane. Due to a distinction existing in their cell walls, bacterial cells generally can be divided into two major categories (Gram-positive and Gram-negative) by using the Gram stain method invented by Hans Christian Gram in 1882 [29]. The difference of chemical and physical properties of cell walls between these two types of bacteria leads to identification by different color appearing after staining (Gram-positive stained with purple while Gram-negative stained with pink).  23  Gram-positive: About 90% of the cell wall is constituted of the thick and compact polymer combined with peptidoglycan and teichoic acid mainly, which is the outmost layer of the cell wall. Therefore the cell wall structure is relatively simple. Gram-negative: The cell wall is thinner and less compact than the one of Gram-positive but more chemically complex. Comparing to the Gram-positive, the layer is composed of only 5~20% peptidoglycan, and is not the outermost membrane. There is another outer membrane surrounding it. The lipid bilayer, protein and lipopolysaccharides (LPS) make up this outer membrane, which is similar to the cytoplasmatic membrane but less permeable. Figure 2-6 shows the basic structure of Gram-positive and Gram-negative cell walls and cytoplasmatic membranes.  Figure 2-6: Cell wall structures of Gram-positive and Gram-negative cells (Source: Dr. Kaiser's Microbiology Home Page).  24  The different compositions of Gram-positive and Gram-negative cell walls give them different electrokinetic properties. The teichoic acids incorporated into the Gram-positive cell wall create a charged network facing outside with high densities like ion-exchange materials [30]. With these open charge groups but excluding the outer membrane of Gram-negative bacteria, Gram-positive bacteria tend to have higher effective conductivities. But there are also exceptions [31] appearing that might be due to differences in strengths of the charged groups in cell walls because the complex LPS is acting like comparable ion-exchange materials [25] in Gram-negative bacteria as well. According to Equation 2-1, not only is the K factor related to the permittivity of the particle hence the conductivity of the particle, but also related to the permittivity of the suspending medium hence the conductivity of the suspending medium. By applying the uniqueness of different cell-wall properties, the manipulation, separation and trapping of cells of interest in a mixture sample can be possibly achieved via the adjustment in conductivity of suspending medium.  2.3 Advantages of Applying Microdielectrophoresis Taking advantage of rapid development of microfluidic technology, DEP is able to be integrated and exploited within a miniaturized scale device. Recently, DEP has been  25  applied as a tool to selectively capture, separate and immobilize bio-particles in the biotechnology field because of its ease of implementation and versatility of purpose. Another advantage of DEP is that it can generate either positive-DEP or negative-DEP or a combination of both [32] for certain upstream processes by adjusting the conductivity of surrounding medium and applied frequency.  2.3.1 Previous Works Overview A DEP-integrated microsystem with PDMS microvalve control was presented by Lagally et al.[33] in 2005 for dielectrophoretic concentration and genetic detection with defined volume (100 nL) of Escherichia coli MC1061 cells. The cells were successfully trapped on interdigitated microelectrodes with positive-DEP force produced by 7 Volts peak-to-peak (Vp-p) and 1 kHz followed by a 30 min genetic detection of 25 cells. Another example of particle trapping was proposed by Gadish and Voldman [34]. They introduced a high-throughput positive-DEP microsystem to concentrate both beads and B. subtilis spores with a chaotic mixer to achieve high-throughput (>100 μL/min) but without a defined volume of the sample controlling by microvalves. Manipulation of detecting targets by using DEP is not limited only to bacterial cells but also for mammalian cells [35], virus particles [36], yeast cells [32, 37] and DNA [38].  26  Apart from utilization of p-DEP, n-DEP also shows itself amenable to manipulation of cells. Due to the repelling force from electrodes affected by n-DEP, particles can be pushed to one side or be caged in a certain area to achieve a trapping or separation. Jaeger et al. [37] cultivated the yeast cells suspended in cell medium by concentrating them in the middle of a channel for several hours through applying n-DEP. Moreover, n-DEP was employed to sort viruses, bacteria, or mammalian cells from complex mixtures by Kim et al. [39]. It was demonstrated that single pass segregation can produce a~1000-fold of concentration of multiple target cell types.  2.3.2 Advantages and Limitations of Positive DEP Although both p-DEP and n-DEP have the ability to capture cells, researchers still have to consider which approach is more appropriate for their experimental purposes after balancing the advantages and disadvantages of these two methods shown in Table 2-1 [22].  27  Table 2-1: ―Comparison of advantages and disadvantages of p-DEP and n-DEP approaches to trapping cells.‖ [22]  In this work, we have selected the p-DEP approach for the trapping process because it can provide a stronger dielectrophoretic force, less heating to minimize thermal damage to cells, simplicity of trapping the cells and simplicity of increasing the force as voltage increases. Moreover, more cells will be trapped by p-DEP than pushing them aside from the electrodes since n-DEP cannot be maximized by turning up the voltage as with positive dielectrophoretic force. The captured cells on the electrodes also offer an easy way for cell lysis with a defined and concentrated sample. While the strongest electric field may more or less facilitate the lysis by electroporation. Compared to the advantages of n-DEP, p-DEP better corresponds to our project goals. Under the condition of adhering on the electrodes with strong electric fields, previous reports [25, 28] show that mammalian cells tend to have irreversible damage when the induced membrane potential is larger than about 1 V with a value of applied electric field  28  exceeding 2 x 105 Vm-1. This damage is caused by the effect of electropermeabilization and electrofusion when the membrane is under electrical stress at low applied frequency. However, bacterial cells have more complex cell wall structures and tend to be more robust to the exterior environment, which has been evidenced by an earlier work presented by Markx et al. [31]. In addition, besides the dielectric properties of the cells, the conductivities of media play an important role on determining the crossover frequency which is used to compare and identify cells with various dielectric properties [40].  Therefore, in order to produce  p-DEP, the suspending media has to be less conductive and polarized than the tested cells, which will lead to biological disruptions to the cells. The effects of suspending buffer on tested bacterial cells are discussed in Chapter 5.  2.3.3 Electrode Design As the fundamental element used for generating either p-DEP or n-DEP, electrodes can be fabricated in a large range of geometries, including interdigitated electrodes, quadrupole electrodes, octopole electrodes, strip electrodes, castellated, and round electrodes [22], and be arrayed in two-dimensional or three-dimensional form for various objectives. At the same time, the difficulty of fabricating microelectrodes must be considered to prevent any modeling or practical problems.  29  Determining which type and what dimension of microelectrodes are suitable for the experimental objectives is the initial step of importance. In order to minimize the thermal and electrochemical disruptions to the cells, it will be beneficial to create a set of electrodes as small as practically possible. Because an operating voltage will have a difference of n3/2 while the dimension of electrodes with any geometries scales down n-fold, one can reduce 1000-fold the operating voltage to create the same DEP force on a particle by cutting down 100-fold of the electrode dimension resulting a 10-fold decrease in the applied electric field. The electrical heating and electrochemical effects are therefore attenuated [25]. Considering the approachability and facility of fabricating the microelectrodes in the UBC cleanroom, the interdigitated electrodes with a size of 20 μm width and spacing is selected after attempting to fabricate electrodes with dimensions of 15, 25 and 40 μm width and spacing. Fabrication processes will be further elucidated in Chapter 3.  2.3.4 Summary In summary, for two or more bioparticles with different dielectric properties, concentration and separation of cells of interest can be achieved by using the difference of polarizability between particles and adjusted suspending media under certain electric field conditions produced by microelectrodes. The polarized particles will move to either the strongest field region or the weakest one with the effects of both dipole moments and  30  electric field. In spite of this, a combination with florescent indictor or utilizing an optical detection platform will be valuable for the manipulation of microorganisms with DEP in effective way. In clinical industry, pathogenic bacteria with similar clinical symptoms are hard to distinguish and recognize rapidly. DEP has the potential ability to tell them apart, working as a promising tool for diagnostic detection.  31  CHAPTER 3: MICROFABRICATION Fabrication techniques involved in micro- and nanotechnology have been under development for several decades. Since the first transistor was invented in 1947, the rapid growth of microfabrication applied in the integrated circuit industry has ignited and microfabrication  has  become  an  indispensable  method  for  microelectronics,  optoelectronics, micromechanics, micro-optics, nanotechnology, micromachines and microfluidics [41]. Microfluidics refers to a multidisciplinary field of manipulating fluid under controlled behaviors at the size scale of 100 microns or smaller. In this section, the fabrication processes of the presented microfluidic device will be discussed. Two major components are included — the polydimethlsiloxane (PDMS) chip containing a single microchannel and the glass substrate deposited with pairs of interdigitated microelectrodes. They are then bonded by exploiting the ultraviolet light ozone cleaner for further experimental testing.  3.1 Techniques and Facilities Device fabrication was accomplished in the class 100 and class 1000 cleanrooms, which have at most 100 particles per cubic meter and 1000 particles per cubic meter  32  respectively, (Figure 3-1) located in The Advanced Materials and Process Engineering Laboratory (AMPEL) at UBC.  Figure 3-1: Cleanroom in AMPEL at UBC. (Source: AMPEL website.)  3.1.1 Photolithography Generally, lithography is the one of the most essential methods playing a role in many minimization processes, and photolithography is the most commonly applied method for transferring desired features printed on masks onto thin layer films.  33  The typical steps included in basic photolithography process are photoresist coating, ultraviolet (UV) exposure, development, etching and photoresist removal. The flow chart is epitomized in Figure 3-2 shown below.  Figure 3-2: Basic photolithography process flow chart. A. Photoresist coating; B. UV exposing; C. Developing; D. Etching.  The photoresist (positive or negative) is spin-coated on a substrate by using a programmable spinner (Laurell WS-400-6NPP-LITE). UV exposure is performed using a  34  Canon Mask Aligner PLA-501F (Figure 3-3) whose UV optics is centered at a wavelength of 400 nm. The etching step follows the photoresist developing process.  Figure 3-3: Canon Mask Aligner (Model: PLA-501F). (Source: AMPEL website.)  3.1.2 Electron-beam (e-beam) Evaporation The e-beam evaporation is one of physical vapor deposition (PVD) methods pertaining to thermal evaporation. Typically, there is a high intensity electron beam (2-20 keV) emitting out and being rotated by a magnetic field under vacuum conditions. The  35  beam will eventually focus back in a graphite crucible filled with metal located in a hearth cooled by running water when the evaporation process is proceeding. The metal will melt when the temperature reaches its melting point and vapours are produced and deposited on the surface of the mounted substrate, with a deposition rate ranging from 50 to 500 nm/min [42]. Figure 3-4 shows a simple schematic of e-beam evaporation.  Figure 3-4: Electron-beam evaporation schematic.  The e-beam evaporation decreases the chance of having source-contamination problems but it might also cause radiation damage and ion damage that ruin the substrate [42]. The e-beam evaporation has been operated in class 1000 cleanroom in AMPEL is a  36  evaporation system with the capability of evaporating four materials by e-beam evaporation and one by thermal evaporation all in one single process (Figure 3-5).  Figure 3-5: E-beam evaporation system. (Source: AMPEL website.)  3.1.3 Plasma-enhanced Chemical Vapor Deposition (PECVD) PECVD is a method commonly used for depositing thin layers on semiconductor wafers. There are two electrodes installed in the reacting chamber generating radio frequency (RF) or direct current (DC) to activate the reacting gases to form plasma for chemical reactions [43]. Figure 3-6 illustrates the EPCVD process.  37  Figure 3-6: PECVD schematic.  The PECVD system located in the cleanroom (Figure 3-7) can be applied for the process of depositing silicon dioxide (SiO2) film by inputting dimethylsilane (DES), nitrous oxide (N2O) and oxygen (O2) into reacting chamber. In this work, PECVD is applied to coat a thin layer of SiO2 on surfaces of the electrode chips.  38  Figure 3-7: PECVD/RIE: Trion. (Source: AMPEL website.)  3.2 Polydimethylsiloxane (PDMS) Channel Chip  3.2.1 Mask Design A flat UV-permeable glass or quartz plate coated with a thin layer of UV-absorbing metal, such as chromium, in 1:1 desired features is traditionally applied as a photomask for photolithography process in microfabrication. This optical glass or quartz mask can create a high resolution with dots per inch (DPI) up to 40000. However, it is relatively expensive as well. Therefore, the photomasks (CAD/Art Services, Inc., OR, USA) used here are comparatively less expensive. They are transparencies printed with black patterns with 20000 DPI resolution.  39  Two simple designs are integrated with dual channels and single channel, respectively, with the same dimension of 3.8 cm length by 1.4 mm width. The dimension is determined to fit the size of certain glass substrate deposited with gold microelectrodes. The channel chips are bonded to electrode substrates with different designs and applied for trapping experiment and separating experiments. All channels are designed with 100 μm height. The dual-channel design can offer a duplicate usage on a single chip, which will facilitate chip fabrication and increase device yield. Smooth semi-circle terminals at both sides of main channels are fabricated to avoid trapping less sample liquid in dead zones. Both main channels with a length of 2.8 cm have two 490 μm-long small sections on the sides connecting to 700 μm radii inlets/outlets. The inlets/outlets contain spider-net excogitations in order to keep a continuous flow traveling into the main channels wherever injecting needles punch through within the circle area. The spider-net excogitations help to divert the sample liquid if injecting needles are not centrally inserted into the inlet/outlet. The single-channel design has the same layout of the dual but with a longer length of 3.08 cm. This 2.8 mm difference is not expected to make any tremendous difference in experimental tests but providing a right alignment to certain designs of the electrode-chips. The mask for fabricating channel chips is shown in Figure 3-8.  40  Figure 3-8: Masks of PDMS channel. A. Dual channel design; B. Single channel design.  3.2.2 Fabrication The molds of presented microchannel chips were manufactured on 4’’ silicon wafers (University Wafer) by structuring an epoxy based negative photoresist—SU-8, which was initially used for the fabrication of advanced semiconductor devices in the microelectronics industry. It has been widely used to provide high aspect ratio images for Micro Electro-Mechanical System (MEMS) for many years.  41  Negative photoresist SU-8 2075 (MicroChem Corp., MA, USA), which is an improved formulation of SU-8, was used to build the molds to delineate channels in PDMS. The main procedures are listed below: 1. Preparation of substrates: A four-inch silicon wafer is first cleaned by a 5:1 mixture of sulphuric acid (H2SO4) and hydrogen peroxide (H2O2) (Piranha wash) for 15 minutes and washed using DI water for 5 minutes. The wafer was blown dry with nitrogen and placed into a 200 Celsius degree (℃) oven for a 20-minute dehydration process to get rid of the extra moisture left on the wafers.  2. Spin-coating: Following removal from the dehydration oven and cooling for 5 minutes to room temperature, the wafer was then secured centrally on the chuck of the Laurell spinner stably and evenly. SU-8 was spun on at 1800 revolutions per minute (rpm) with 300 rpm/s ramping for 40 seconds to achieve uniform 100 μm-thick layers of photoresist on the surface of wafers.  3. Soft bake: The SU-8 coated wafer was placed on a hotplate for pre-UV-exposure baking to cure the photoresist. The baking time was 4 minutes at 65 ℃ initially and ramped up to 95 ℃ for 22 minutes followed by 2 more minutes at 65 ℃. Ramping up/down as gently as possible helped avoid bubble formation.  4. UV-Exposure: The baked wafer was cooled down for 2 minutes and exposed to UV light in the Canon mask aligner with the photomask in hard contact for 180 seconds in  42  an intermittent manor, meaning the exposure time was divided into 3 times with 60-second exposure following with a 30-second break. This prevented the photoresist from overheating.  5. Post bake: The UV-exposed wafer was returned to the hotplate. The baking time was 4 minutes at 65 ℃ and 10 minutes at 95 ℃, then ramped down to 65 ℃ for 2 minutes. In the same manner as the soft bake, mild temperature ramping helped avoid damage to the photoresist film.  6. Development: The wafer was cooled down and then immersed in SU-8 developer (MicroChem Corp., MA, USA) for 3 minutes with continuous agitation. The wafer was removed and flushed with isopropanol (IPA) for half the developing time to remove traces of unexposed SU-8. The wafer was blown dry and backer-immersed into developer to continue the development until the features were visibly clear. The wafer was rinsed with IPA and blown dry and inspected under a microscope.  7. PDMS chips: The PDMS was prepared using a Sylgard® 184 Silicone Elastomer Kit (Dow Corning Corporation, MI, USA) with a 10:1 ratio of elastomer to curing agent and blended thoroughly for several minutes. The viscous mixture was poured slowly onto a silicon wafer mold pre-placed on a 5’’ aluminum plate to achieve an even surface, ensuring that the PDMS covered the whole surface of the mold. The substrate was placed in a desiccator for degassing and checked every 5 minutes until all the  43  bubbles were eliminated. The degassed PDMS mold was placed into a pre-heating oven at 80 ℃ and baked for 120 minutes. The mold was cooled to room temperature and the PDMS channels were peeled from the silicon mold. The cured PDMS channels were placed back into the oven for baking for another 12 hours. The PDMS chip was then be diced into appropriate size for further use.  3.3 Interdigitated Microelectrodes Chip  3.3.1 Mask Design The mask design for the interdigitated microelectrodes was designed to insert into a custom-design surface-mount (SMT) board (Figure 3-9) designed by Arash Takshi. Not only does the board function as an experimental platform, but it also serves as an electrical connection between the function generator and microelectrode chip placed on it. The SMT board contains two additional acrylic clamps of 47 mm length and 11 mm width screwed on both sides of the observed window. A connecter with 25 pins is soldered on each of these clamps to connect to the microfluidic chip. Each of three pins are designed to link to 16 defined BNC connectors soldered on the side of the SMT board. The circuit board is incorporates the capability of signal amplification but this was not being used in this work.  44  Figure 3-9: Costumed-design SMT board and strip connector soldered on acrylic clamp.  The schematic mask (CAD/Art Services, Inc., OR, USA) design for fabrication of the interdigitated microelectrodes is shown in Figure 3-10. Both electrode chips were 4 cm long and 3 cm wide that allowed the PDMS channel chips to be bonded appropriately. With the purpose of observing the experiments processed on chip in real time, the juxtaposed sets of electrodes for the dual channel chip in the first mask pattern (Figure 3-10A) contained 276 pairs of microelectrodes of 20 μm width and spacing within a rectangle region of 22.06 mm length and 7.4 mm width. The mask design also incorporates  45  two bus strips of 22.02 mm length and 250 μm width on both sides of the electrodes to connect them together. Another two bus strips of 300 μm width were connected to two adjacent pads at a distance of 400 μm. With the aim of ensuring these two pads obtaining a good touch-connection with those pins on one of two connectors screwed on the SMT board, each pin was 2 mm long and about 0.5 mm wide with 0.5 mm spacing between the two. As three pins have to link to one pad, the width and length are recommended to be 28 mm and 25 mm, respectively. The second mask (Figure 3-10B) has a similar design to the first mask (Figure 3-10A). There are two separated sets of electrodes with 120 μm distance in series. Each of them contains 188 pairs of electrodes of 2 mm total length. Two measurement pads on each side were added to facilitate the measurement of the on-chip-signal when it was connected in the SMT board.  46  Figure 3-10: Masks of electrodes. A. Juxtaposed sets of electrodes for dual channel design; B. Tandem sets of electrodes for single channel design.  47  3.3.2 Fabrication Four inch borofloat (Precision Glass & Optics, CA, USA) glass wafers were used as substrates and diced into chips of certain sizes required for experiments. The features were patterned by spin-coating MEGAPOSITTMSPRTM220-7.0 photoresist (MicroChem Corp., MA, USA) using a photolithography technique similar to the PDMS channel chip fabrication described above. Unlike SU-8 2075, the SPR 220-7.0 is a positive photoresist, which will dissolve into the developer when exposing to the UV light. This specialty helped create a sacrificial mask from lifting off by the etchant, described in detail below. 1. Substrate-cleaning: A 4’’ glass wafer was placed into beakers filled with acetone and sonicated for 10 minutes, then transferred into IPA for another 10 minute sonication cycle. The wafer was rinsed with DI water and placed into a 200 ℃ oven for 20 minutes dehydration following blow drying with nitrogen.  2. Metal deposition: The cleaned wafer was mounted face-down in a sample holder in the e-beam evaporator. Crucibles filled with Chromium (Cr) or Gold (Au) were successively placed into the switchable chuck embedded in the hearth. Evaporation was performed at case pressures lower than 5.0X10-6 Torr,. The e-beam was started and the power gradually increased until the metal was ready to evaporate. A 20 nm thick Cr layer followed by a 200 nm Au layer was deposited on the substrate surface.  48  Deposition rates around 3.0 to 4.0 angstroms (Å) per second were maintained to create a relatively uniform metal layer on the substrates.  3. Piranha wash:The Cr/Au-coated glass wafer was cleaned with piranha solution for 10 minutes, followed by immersion in a crystallizing dish containing DI water for 5 minutes and then rinsed with DI water thoroughly. The wafer was blown dry and dehydrated in a 200 ℃ oven for 20 minutes.  4. Hexamethyldisilazane (HMDS) priming: HMDS is a common adhesion promoter used to generate a hydrophobic interface on the surfaces of substrate that will readily enhance the adhesion between photoresist and wafers[42]. The wafer was cooled and centered on the chuck of a spinner. A few drops of HMDS were manually pipetted on the wafer surface and the wafer was then spun at 4000rpm for 40 seconds to dry.  5. Photoresist-coating: SPR-7.0 photoresist was poured onto the middle of HMDS-treated wafer and spun for 10 seconds for 500 rpm with an acceleration of 300rpm following with 5300 rpm with an acceleration of 600 rpm for 40 seconds to achieve a thickness of about 5.5 μm. 6. Soft bake: The photoresist-coated wafer was baked on a hotplate at 115 ℃ for 120 seconds and transferred onto a wipe to cool for 2 minutes before UV exposure.  49  7. UV exposure: The wafer was exposed in a UV aligner for 65 seconds to achieve an exposure does of 150 mJ/cm2 or a light integra of 12.  8. Post bake: The wafer was stored in a glass Petri dish for at least 30 hours before post exposure bake. Experience showed that the longer delay time results in better pattern resolution. Post bake was performed on a hotplate at 110 ℃ for 120 seconds. 9. Development: The wafer was cooled and immersed in MFTM-24A developer (MicroChem Corp., MA, USA) for approximately 3 minutes with continuous agitation. The wafer was removed following half of the developing time, rinsed with DI water, and blown dry. The wafer was then replaced in the developer and agitated for another one and a half minutes. The wafer was then removed and rinsed with DI water again. Following washing, the wafer was blown dry and observed in a microscope to determine if more developing time was needed.  10. Etching: The photoresist left on the surface of the wafer was used as a sacrificial mask to protect the electrode pattern from etching by the metal etchants. The developed wafer was placed into Au etchant for 72 seconds with agitation, and the 200nm thick gold layer was removed at a etching rate of 28 Å per second. The wafer was washed in DI water and blown dry before immersion in Cr etchant for 15 seconds to remove the Cr layer. The wafer was again washed in DI water and blown dry. The remaining  50  photoresist was removed by immersing the wafer in acetone for few seconds. Finally, the wafer was rinsed with IPA and DI water and blown dry.  11. Plasma-enhanced chemical vapor deposition (PECVD) treatment: The wafer was diced before deposition of a SiO2 film by exploiting PECVD. The SiO2 film was intended to lower the non-specific binding of unwanted particles to the electrode surface. Deposition was conducted at a pressure of 450 millitorr and a temperature of 190 ℃, using a DES flow rate of 4 cubic centimeters per minute (SCCM) and a N2O flow rate of 71 SCCM at a power of 100 Watts.  3.4 Device Packaging  3.4.1 PDMS-glass Bonding PDMS-glass bonding can be accomplished in a reversible way, which is derived from a weak van der Waals force less than five pound per square inch (psi) [44] and an irreversible way, such as oxygen plasma bonding, which changes the surface chemistry of the PDMS and glass [44, 45]. Instead of using oxygen plasma bonding method, a UV/ozone cleaner (Model 42, Jelight Company, Inc., CA, USA) was used to modify the surfaces of PDMS and glass for bonding at room temperature.  51  The UV lamp had an emission wavelength (λ) from 185 to 284 nm. Ozone is produced with oxygen in an atmospheric reaction with the UV light at a wave length of 185 nm. Combining with the ozone, a concurrent excitation of organic molecules will happen when λ=284 nm [46], as shown in Figure 3-11.  Figure 3-11: Ozone reaction mechanism.  As an organic molecule, PDMS is a compound formed with lots of Si-CH3 methyl groups that make it hydrophobic. Modification is necessary to proceed to change the surface chemistry to be temporarily hydrophilic such that it can bond to the glass substrate. O3, a strong oxidant, can oxidize the organic fragments of PDMS to cause a transformation from -O-Si-CH3 groups to -O-Si-O- network with byproducts of CO2 and H2O, as shown in Figure 3-12.  52  Figure 3-12: PDMS surface chemical structure transformation.  The proportion of the -O-Si-O- groups to the non-reacted -O-Si-CH3 groups depends on the period of UV treatment, the duration of which ranges from 30 to 230 minutes [47]. UV/ozone also provides a method of reducing the contamination on the surface of glass substrate and creates hydroxyl (-OH) groups on it. This will allow the hydrophilic PDMS surface easily to bond to the glass substrate. In this work, the electrode chip was prepared by basic solvent cleaning (acetone for 10 minutes/IPA for 10 minutes/DI water) then blown dry. The PDMS channel chip was cleaned using Magic™ Tape (Scotch® Brand, MN, USA) to remove as much as dust visually resting on the surface. The inlet and outlet were punched with a hole-punch of 0.5 mm diameter (Harris Uni-CoreTM, CA, USA) creating holes. Both chips were placed into the ozone/UV cleaner and reacted for 5 minutes. The PDMS chip was aligned with the electrodes, and manual pressure was applied to remove bubbles and ensure a good sealing. The assembled chip (Figure 3-13) was left to rest overnight to create a strong bond.  53  Figure 3-13:  Fully assembled DEP chips with electrode substrates bonded to PDMS channels.  3.4.2 Chip Assembly Device packaging is an essential step to build up a link between a micro-device and the macro-scale world. Here, the SMT board mentioned in section 3.3.1 functioned as the link between the two. External electrical access to the device was accomplished through BNC connectors and the two strip clamps pressing down on the device. While fluidic access is provided via the flexible plastic tubing (TYGON®, OH, USA) inserting into the inlet and outlet. The screwable clamps help fix the device chip at the correct position on the  54  board once the pins have a good alignment and contact with the connecting pads on the device, as shown in Figure 3-14.  Figure 3-14: Assembled and packaged DEP chip located within the SMT board. Fluidic tubing and BNC connectors are visible.  The assembled device can be switched to another chip incorporated with the other electrode geometry when necessary. The assembled chip demonstrates good fluidic performance when injecting samples because of the good sealing endowed by the irreversible bonding. However, the chip is not durable if there are leaks from the inlet, outlet, and the side or the electrodes become shorted.  55  CHAPTER 4: MATERIALS AND METHODS This chapter describes the materials used for the experimental methods involving the DEP microfluidic system, and concludes with a proposed method to evaluate the performance of the microsystem.  4.1 Experimental Materials Pathogenic bacteria are the primary test targets in this work, Salmonella typhimurium (S. typhimurium), Campylobacter jejuni (C. jejuni), Escherichia coli (E. coli) and Mycobacterium smegmatis (M. smegmatis) were trapped and concentrated in the microconcentrator. Polystyrene microspheres were also trapped as initial testing was conducted. Phosphate buffered saline (PBS), HEPES buffer (4-(2-hydroxyethyl)-1piperazineethanesulfonic acid) and cerebrospinal fluids (CSF) were applied as suspending buffers to introduce the bacterial cells into the PDMS channel for capturing tests.  4.1.1 Suspending Fluid Media All the media preparations were performed in a classⅡ, A2 biological safety cabinet (Model 1284 Series, Thermo Scientific, MA, USA).  56  PBS Buffer 10 mM PBS buffer was made by dissolving PBS tablets (1 tablet/ 100ml) (Invitrogen, 3002) into DI water following with an autoclaving step. Sterilized PBS with a conductivity of 17.6 mS/cm was then adjusted by adding DI water to meet desired conductivities for various purposes of trapping. The required conductivity, as low as 40 µS/cm, is very dilute and close to that of deionized water, and whose density is approximately 1000 kg/m3. With the purpose of sustaining the particles floating in sample filling in the syringe during proceeding experiments, the specific gravity or relative density of the diluted PBS buffer was amended by adding 10% dextrose (D16-3, Fisher Scientific). However, dissolved dextrose was found to change the conductivity of the buffer, so a final conductivity correction was necessary before use. Modified buffer was filtered by Millex®-GV 0.22-µm-filter-unit (Millipore Corp., MA, USA) and stored in sterilized containers at 4 ℃. HEPES Buffer HEPES is one of twelve buffering agents (Good's buffers) with a zwitterionic property including both positive and negative charge. (See Figure 4-1) Because of its better performance of maintaining physiological pH, it is generally applied in cell cultures. The value of dissociation constants (Kd) will drop off with declining temperature, which is also better for the maintenance of physical structure and function of enzymes at low temperatures [48].  57  10 mM HEPES was prepared by adding 2.6 g HEPES sodium salt (75277-39-3, SIGMA®) into 1000 ml DI water followed with a sterilized process. The conductivity of 10 mM HEPES is 0.98 mS/cm, thus DI water dilution is required to achieve the correct conductivities for bacterial cell trapping. However, HEPES buffer shows a higher concentration than PBS buffer when diluting to the same conductivities with DI water. This is due to the zwitterions in HEPES, which will help to retain the physiological concentration in the buffer at low conductivities. Therefore, HEPES is the primary buffer applied as the suspending media in most of the experiments. The working buffer was finalized by specific gravity modifications with 10% dextrose and filtering processes as what have been described for PBS, then stored at 4 ℃.  Figure 4-1: Chemical structure of HEPES.  Cerebrospinal fluid (CSF) Unalterated and pooled CSF samples were kindly provided by Dr. Patrick Tang at the BC Centre for Disease Control (BCCDC). CSF is a watery fluid established around and inside the brain and spinal cord. It acts like a liquid cushion that the brain floats in,  58  reducing the weight of the brain by 97%. Mechanical injuries to the central nervous system can be mechanically protected and cushioned by CSF. In addition, the brain is also nourished by the CSF even though rich blood supply exists. Neurons generate waste products and CSF can assist to get rid of them. CSF has more sodium and chloride ions and less protein but it is still similar to the blood plasma where it arises[49]. It is aimed at verifying that bacterial cells are able to be trapped in bodily fluids via a simple dilution process. CSF is selected because various situations affecting the central nervous system (CNS) can be confirmed by detecting a sample of CSF and the results are more easily achieved than other bodily samples, such as urine or blood, which are easier to be collected. For example, meningitis caused by infections and inflammation in the meninges can be identified through a CSF examination. Therefore, CSF is an effective tool to expedite the process of detecting many diseases in diagnostic medicine. The CSF samples used in this study were pathogen-free. The original sample of CSF had a conductivity of 0.94 mS/cm, which is too high to produce p-DEP. Subsequently, it has to be diluted by sterile DI water to reach preferred conductivities for trapping.  4.1.2 Polystyrene Microsphere Biological particles, bacterial cells of single species for instance, have diverse morphologies and dimensions even within one sample. With the aim of characterizing the  59  device to see if it can reliably generate p-DEP for trapping bacterial cells, beads with well-defined and uniform size similar to bacterial cells are initially tested according to the cell model mentioned in Chapter 2. The beads were 1.90-µm-diameter polystyrene microsphere with Dragon Green fluorescent dye (excitation: 480 nm, emission: 520 nm) (Bangs Laboratories, Inc.) showing insignificant photobleaching. The sizes of these beads are small enough and close to the size of the bacteria of interest in these studies, such as E. coli. The fluorescent property of the beads can help to observe the trapping conditions in the channel microscopically.  4.1.3 Salmonella typhimurium S. typhimurium (Figure 4-2) is Gram-negative bacterium of rod shape of 2 to 5 µm long and around 0.7 to 1.5 µm in diameter.  60  Figure 4-2: Electron microscope image of S. typhimurium. (Source: Volker Brinkmann, Max Planck Institute for Infection Biology, Berlin, Germany.)  A Green fluorescent protein (GFP)-expressing strain of S. typhimurium SL1344 (hisG rpsL xyl) was kindly supplied by Dr. Brett Finlay’s lab at the Michael Smith Laboratories (MSL), UBC. S. typhimurium cells exploited in this work were cultured in Luria-Bertani (LB) broth (SIGMA®,L7658-1KG) and LB agar (SIGMA®, L7533-1KG) with a dosage of 10 μg/ml tetracycline (Fluka, 87128). To start a new culture, a loop full of culture was removed from a frozen stock stored in cryogenic vials (NALGENE®, NY) and streaked out onto a Petri dish (Fisher Scientific, NH) with LB agar. The plates were placed upside-down into an incubator (Model SymphonyTM 5.3A, VWR®, PA, USA) at 37 ℃ overnight. Single colonies growing on the plate were picked and placed into a culture tube (SARSTEDT, 62.515.006) containing 5 ml  61  LB broth. The tubes were slanted in the rack of a shaking incubator (Modeel 435, Thermo Electron Corporation, MA, USA) and incubated at 215 rpm at 37 ℃ overnight. The next day, the cells were subcultured with fresh media by diluting an aliquot 1:50 of the overnight culture into fresh LB broth, then incubating for another two hour and a half period in the shaking incubator at 37 ℃ for further use.  4.1.4 Campylobacter jejuni C. jejuni 81-176 with GFP expression was obtained from Finlay’s Lab at MSL, UBC. C. jejuni (Figure 4-3) has a spiral bacillus shape and single flagellum at both tips. It is also a Gram-negative microaerophilic bacterium, usually requiring an environment with 5% oxygen, 10% carbon dioxide and 85% nitrogen to survive.  62  Figure 4-3: Electron microscope image of Campylobacter jejuni. (Source: Virginia-Maryland Regional College of Veterinary Medicine, Blacksburg, Virginia.)  Initial culture were obtained by transferring 20 μl of slurry from a frozen stock thawed on ice for 5 minutes, and spotting into the middle of an agar plate made from Brain Heart Infusion (BHI) (DifcoTM, 241830) supplemented with 5% newborn calf serum (NCS) (Invitrogen Corporation, 16010-159), or Mueller Hinton agar (DifcoTM, 225250). For selection purposes, vancomycin (Sigma®, 1404-93-9) of 10 μg/ml and trimethoprim (Sigma®, 738-70-5) of 5 μg/ml were added to the agar plates. Once the liquid patch on the plate dried, the plates were placed upside-down in a Campy pouch system (BD GasPakTM EZ, 260685) (for incubating small amounts of plates) or a Campy container system (BD GasPakTM EZ, 260680) (for incubating large amounts of plates) with a gas-creating pouch for creating a microaerophilic environment for culture. The entire pouch or container was placed into an incubator overnight at 37 ℃. Confluent  63  cell patches were observed on the plates the next day, and a sterile inoculating loop was next used to spread out the patch on the plate. The plates were placed back into the microaerophilic pouch or container system and underwent another overnight incubation at 37 ℃. This step resulted in more cells propagating on the plate. A full loop of cells was retrieved from the plates and resuspended in 10 mM PBS and seeded into a culture tube containing BHI broth (DifcoTM, 237500) with 5% NCS or Mueller Hinton broth (DifcoTM, 211443). Dosages of 10 μg/ml vancomycin and 5 μg/ml trimethoprim were added. The culture tubes were placed into a sealable jar with gas creating packs and placed into a shaking incubator at 37 ℃ at 90 rpm. This broth culture was then subcultured into fresh broth for a further 16 hour growth and was then used for experiments.  4.1.5 Escherichia coli E. coli was used for trapping detection due to its harmlessness and availability as it is normally a research tool for amplification of the amount of plasmids. Generally, E. coli cells are rod-shaped (Figure 4-4), and are Gram-negative with dimensions of about 2 μm long and 0.5 μm in diameter. The test strain of red fluorescent protein (RFP) expressing - E. coli DH5 (fhuA2 Δ (argF-lacZ) U169 phoA glnV44 Φ80 Δ (lacZ) M15 gyrA96 recA1  64  relA1 endA1 thi-1 hsdR17) was created by the 2009 UBC international Genetically Engineered Machines (iGEM) team.  Figure 4-4: Electron microscope image of E. coli. (Source: Rocky Mountain Laboratories, NIAID, NIH.)  Similar to the culture of S. typhimurium, E. coli culture began with inoculation from frozen stock onto a LB agar plate but mixed with ampicillin (SIGMA®, 69-52-3) at 100 μg/ml, then incubated at 37 ℃ overnight. Single colonies were picked and placed into LB broth with a dosage of 100 μg/ml ampicillin. 100 μl of the overnight culture was then transferred into 5 ml flesh LB broth (1:50) following with a two and a half hour shaking subculture at 37 ℃ for further use.  65  4.1.6 Mycobacterium smegmatis GFP-expressing M. smegmatis (ATCC 19420) was kindly provided by Dr. Patrick Chan at the Tuberculosis Laboratory at the BCCDC. M. smegmatis (Figure 4-5) is a Gram-positive and rod-shaped species with a size of 2 ~ 4 um in length and 0.2 ~ 0.5 um in width. It is also an acid-fast bacterium like other Mycobacteria because it resists to be decolorized by the dilute acids during the gram staining processes which makes it seem to be Gram-negative species, although its cell wall structure is considered to be most similar to Gram-positive species [50].  Figure 4-5: Electron microscope image of M. smegmatis. (Source: Molecular Microbiology, Victoria University of Wellington, New Zealand.)  The difficulty of identifying M. smegmatis or other Mycobacteria by using standard Gram staining methods is due to the special properties of their cell structures diverging  66  from most Gram-positive bacteria. They are characterized as having thick and waxy-coating cell walls and compact inner cell membranes presenting hydrophobicity and low permeability. These properties are due to a long, branched fatty acid, mycolic acid, forming a major part of the cell wall. The unique structural features and the relatively slow growth rate compared to most other bacteria render Mycobacteria, including M.smegmatis, to possess the inherent capability of low response to many antibiotics [51, 52]. Contrasting with several harmful Mycobacteria, such as M. Tuberculosis, M. smegmatis is recognized as a safe and non-pathogenic microorganism widely found in the soil, water, and plants. Even though reports of diseases causing by M. smegmatis have been reported [55, 56] M. smegmatis is still safer and easier to work with since it is quicker growing and is classified as requiring the lower biosafety security level BSL-1. Moreover, M. smegmatis shares most genetic information [57] and most importantly cell structural features with M. Tuberculosis [58], it is relatively ideal to serve as a model for researching M. Tuberculosis especially in the DEP-trapping characterization here. Culturing M. smegmatis was performed by taking a full loop of culture from a frozen stock and streaking onto Middlebrook 7H11 agar plates (Bio-Media Unlimited Ltd.,ON, CA). The plates were incubated at 37 ℃ for 48 hour and cells grew in confluent lines on the plate. An inoculating loop was used to scrape cells and relocate them into Middlebrook 7H9 broth (Bio-Media Unlimited Ltd., ON, CA). Tubes containing culture media and cells were then placed into a shaking incubator set to 100 rpm and incubated for 24 hours at  67  37 ℃. The overnight broth culture was pipetted into fresh 7H9 broth diluted 1:50 and incubated for another 18 hour at 37 ℃.  4.2 Testing Particles Preparation and Storage  4.2.1 Non-biological Particles Testing beads were prepared by tenfold serial dilution from the stock of 6.6325×10 9 microspheres/ml. The final concentration was made to be 6.6325×105 microspheres/ml in 10 mM PBS buffer. The beads were washed twice with diluted PBS of desired conductivity by using a centrifuge at 6000 rpm for 5 minutes. The microspheres were stored in a dark bottle containing 10 mM PBS and placed into a refrigerator at 4 ℃.  4.2.2 Biological Particles Cells at exponential phase were centrifuged down at 8000 rpm for 5 minutes. The supernatant was discarded and the pellet resuspended in 10 mM PBS or HEPES. This procedure was repeated three times, with a final resuspension in buffer of the desired experimental conductivity and 5% dextrose.  68  Bacterial cells were all stored in a -80℃ freezer (Thermo Scientific, MA, USA). For S. typhimurium and E. coli, an overnight broth culture proliferated from a single colony was washed twice with fresh LB broth then resuspended in storage medium composed of LB and 15% glycerol (SIGMA®, 56-81-5). The suspension was transferred to a cryogenic vial locating into a cryogenic box and stored at -80 ℃. For C. jejuni and M. smegmatis, similar procedures were performed as with S. typhimurium and E. coli, but the storage media was changed to BHI or Mueller Hinton with 15% glycerol for C. jejuni and 7H9 mixed with 50% glycerol and 20% Tween® 20 (SIGMA®, 9005-64-5) [59].  4.3 Measurement Apparatuses Measurement tools for preparing the buffer, quantitatively determining the testing particles and optically inspecting the operation experimental testing are described in this section.  4.3.1 Conductivity Meter The conductivities of media fluid are fundamental to the performance of DEP trapping. A fast and relatively rapid way to measure the buffer conductivity is necessary. A handheld conductivity meter (Twin Cond B-173, HORIBA, JAP) was used to evaluate the  69  effective conductivities of the suspending media in a fast manner. The meter was calibrated by a known buffer of 1.44 mS/cm before every set of measurements, but the detecting well of the meter was briefly washed with DI water after every measurement.  4.3.2 UV Spectrophotometer Optical assessment is valuable for quantifying the concentration of input or output samples. Concentration difference between the input and output sample is an important reference to define the performance of the trapping efficiency in this study. Due to a limitation of the volume of the test sample we used for experiments, a UVette (Eppendorf, NY, USA) with small volumes (≥50 µl) was used to contain the sample to be measured on a BioPhotometer (Eppendorf, NY, USA). The measured absorbance was established at wavelength of 600 nm for most of the biological particles, including bacterial cells. With the aim of keeping the values of OD600 within linear range according to the calibration curve showed in Chapter 5, dilution was needed to reduce the concentration of sample. Furthermore, to create standard curves for different tested bacterial species, series of different tested cell suspensions of known concentrations were measured with a BioPhotometer at 600 nm wavelength. The results of the detected concentrations were  70  acquired from counting colony-forming unit (CFU) /ml on plates. Graphs are plotted as CFU/ml versus values of absorbance that will be shown in Chapter 5.  4.3.3 Fluorescence Microscope Another essential instrument for optical inspection is necessary to characterize the trapping performance on chip. A Nikon fluorescence microscope (Model ECLIPSE Ti, Nikon Instruments Inc. NY, USA) was applied for visual observation, and a QImaging CCD digital camera (MicroPublisher 5.0 RTV, QImaging, BC, CA) was employed to acquire images in real time. Depending on the fluorescent properties of the observed particles, different filters assembled in the Nikon microscope were switched to meet the purposes of visual inspection. Since there are only two types of fluorescence used in this work, two filters, 49002 ET - GFP (FITC/Cy2) (Chroma Technology Corp.,VT, USA) and 49005 ET DSRed (TRITC/Cy3) (Chroma Technology Corp.,VT, USA), were applied, whose optical properties are listed in Table 4-1. Based on the values of excitation and emission shown below, GFP - expressing bacteria and dragon green embedded microspheres are able to be visualized by ET - GFP (FITC/Cy2), while ET - DSRed (TRITC/Cy3) is appropriate for observing RFP - expressing ones.  71  Table 4-1: Spectra of fluorescent particles and microscope filters. Excitation (nm)  Emission (nm)  ET - GFP (FITC/Cy2)  470  525  ET - DSRed (TRITC/Cy3)  545  620  Beads with Dragon Green Fluorescent Dye  480  520  GFP-expressing S. typhimurium, C.jejuni  470  510  584  607  and M.smegmatis RFP- expressing E. coli  In general, standard operation procedures for operating bacteria have to be strictly followed, even though the testing microorganisms are BSL-1 or BSL-2 bacteria, which are relatively safe to work with. On the other hand, the preparation of suspending buffer was conducted in a Class II, A2 biological safety cabinet in order to avoid unnecessary contaminations that may be possibly affect the DEP responses of the particles as well as the trapping performance.  4.4 Experimental Setup A simple microfluidic system incorporated with the functionality of generating dielelctrophoresis is presented as a concentrator for trapping microorganisms of interest. The experimental setup is illustrated in Figure 4-6.  72  Figure 4-6: Experimental setup.  All tested samples were carried by 1 ml disposable syringe (BD™, NJ, USA) and pressure-driven by a single channel syringe pump (HARVARD APPARATUS, MA, USA). Depending on test purposes, p-DEP force is induced by applied electric fields that are produced by one or two synthesized function generators (SFG-1000 series, GW Instek, CA, USA). Furthermore, pre-set voltages on the function generators were decreased when samples were transferred into the channel. Therefore, a digital storage oscilloscope (GDS-2000 Series, GW Instek, CA, USA) was used to attain the real signals present due to  73  voltage drops across the chip and slight adjustments were made to meet the experimental requirements. As mentioned above, performance of the microconcentrator was monitored by a Nikon fluorescent microscope and images of trapping status shown on the display were taken using a CCD camera. There are only one inlet and one outlet on chip, which are connected by modified 23G needles (BD™, NJ, USA) with right angles as adaptors extending out via TYGON® tubing. The inlet side was hooked to the blunt needle equipped on the syringe while the outlet led to a microcentrifuge tube used for collecting samples.  4.5 Experimental Procedures Experiments were all run on the microsystem with a similar protocol for testing, including material preparation and instrument examination described previously. Nevertheless, procedures have modified as appropriate and these modifications are described below.  4.5.1 Concentration Efficiency Evaluation The device was set up as illustrated above prior to operating an experiment, as well as the preparation of testing samples. The input sample was contained in a syringe at a known  74  concentration, which was calibrated according to the specific standard curve of the tested particle. Dilutions with buffers (diluted PBS, HEPES or CSF) were then made to reach the desired concentration for trapping. A more accurate concentration (CFU/ml) of input sample was further established by counting colony results on plates. Previous to injecting the sample, air bubbles in the syringe cylinder were eliminated. A small amount of negative suspending buffer containing 1% Tween® 20 was flushed at 1 ml/hr into the channel to remove the air and to coat the surface area. Next, the desired voltage and frequency were applied to generate p-DEP force on the electrodes and sample was introduced into the channel at 150 μL/hr for 10 minutes. Captured particles tended to stick on the electrodes due to the p-DEP while uncaptured ones in suspending buffer were eluted from the chip and flowed into a 500-μl microcentrifuge tube pre-filled with 50 μl of 10 mM PBS or HEPES buffer. The pre-filled buffer of higher concentration in the collected tube is to help preventing the biological particles from staying too long in the diluted suspending media and dying quickly once they come out from the outlet. At the same time, optical inspection of trapping performance was accomplished through the Nikon fluorescence microscope and images and videos were recorded by CCD camera. The collected tube was capped and removed to avoid any evaporation and contamination. It was then stored in an ice bucket for plating to obtain quantitative  75  information comparing its result to that of the input sample. To release the captured particles, the applied p-DEP signal was removed and 1 ml clear buffer containing 1% Tween® at a higher flow rate (1 ml/hr) was introduced to flush cells from the channel. Another 500 μl 70% ethanol was next flushed into the channel and let sit for 10 minutes to sterilize and remove any residues on the electrodes. Waste was collected from the outlet using another tube. The input and output samples were serially diluted and cultured on agar plates. Colonies were counted within one or two days depending on the species of the cells. The same experiment for each tested particle population was conducted at least three times and then results of comparing both input and output samples were analyzed and are presented in Chapter 5.  4.5.2 Particle Effective Conductivity Assessment Similar to those processes for evaluation of trapping efficiency, the concentration of the input sample was determined by BioPhotometer before injecting through the microdevice. Further, the absorbance value of the output sample was obtained after flowing through the channel under induced electric field for 10 minutes. Due to the high limit of detection of the BioPhotometer measurement, input concentrations were higher than those used in experiments of trapping efficiency  76  assessment in order to acquire a relatively higher absorbance value from the output sample. These processes were repeated, employing a series of different conductivities of suspending buffer ranging from 20 μS/cm to 680 μS/cm. Control experiments were also performed by flowing input sample into the channel without applying DEP and collecting the sample after 10 minutes. Three parallel results of OD600 absorbance collected from these experiments were analyzed and are elucidated in Chapter 5.  4.5.3 Mixed Sample Separation Experiments were performed with a similar protocol to that described above except that two species of bacterial cells were used instead of just one. The experimental chip design was also changed to the one integrated with two sets of electrodes. Experimental suspending buffers of desired conductivity and different applied frequencies were selected in terms of the analysis of effective conductivities of both species of interest, but the same voltage was applied to both sets of electrodes on the chip. Optical monitoring was used as before getting to assess the separation qualitatively by taking images and videos.  77  CHAPTER 5: RESULTS AND DISCUSSION The demonstration of the functionality and characterization of the microsystem are presented through the data obtained and analyzed through microscopy and quantitative measurement by comparing both input and output sample concentrations.  5.1 Standard Curves Standard curves of bacteria were used to translate the OD600 absorbance values to the input or output concentration. Bacterial cells used for creating the curves were collected at exponential growth phase, and the same samples were used for all the experiments. Measured samples were made by serial dilutions and measured by a BioPhotometer at 600 nm wavelength.  5.1.1 Salmonella typhimurium A determination of logarithmic phase for S. typhimurium cultured in LB media with tetracycline was made before creating the standard curve. Single colonies were picked and placed into culture tubes filling with LB plus tetracycline for overnight culture followed by a 200-minute subculture. Sampling and plating were conducted four times every 30  78  minutes, then 80 minutes for the last time. Plate counting data were analyzed and used to plot a growth curve of concentration (CFU/ml) versus culture time presented in Figure 5-1.  Figure 5-1: Sample growth curve of S. typhmurium growing in LB broth with tetracycline for 200 minutes.  Figure 5-1 shows the change of concentration (CFU/ml) of S. typhimurium growing in LB broth with tetracycline as a function of time. As S. typhimurium have an exponential increase after 150 minutes, a 2.5-hour (180-minute) culture time was chosen for collecting S. typhimurium cells at logarithmic phase.  79  A 2.5-hour LB broth subculture from an overnight culture at 37 ℃ was collected and prepared for OD600 measurement. A standard curve of concentration (CFU/ml) against the values of absorbance obtained at a wavelength of 600 nm was created and is shown in Figure 5-2.  Figure 5-2: Sample calibration curve of S. typhmurium growing in LB broth with tetracycline.  80  5.1.2 Escherichia coli Applying similar procedures to those used for acquiring a growth curve for S. typhimurium, a growth curve of E. coli growing in LB broth with ampicillin was plotted as Figure 5-3.  Figure 5-3: Sample growth curve of E. coli growing in LB broth with ampicillin for 200 minutes.  81  The E. coli has entered into logarithmic phase after 60 minutes and maintains this trend until the end of the experiment. Thus 2.5-hour (180 minutes) is also determined to be the total culture time needed for collecting the E. coli cells at exponential phase. A subculture of E. coli in LB broth with ampicillin for 2.5 hours is collected and prepared for OD600 measurement. The standard curve of E. coli is plotted, as shown in Figure 5-4.  Figure 5-4: Calibration curve of E. coli growing in LB broth with ampicillin. (Error bars were presented but too small to perceive.)  82  5.1.3 Campylobacter jejuni Campylobacter is a slow grower on all media, with a doubling time of approximately 3 hours depending on the strain [60]. In this work, 16-hour culture time was designed for creating the standard curve of C. jejuni at exponential growing phase [60, 61]. Samples cultured on two types of media — BHI broth plus 5 % NBS and Mueller-Hinton broth (both with vancomycin and trimethoprim) for 16 hours were collected and washed twice with 10 mM PBS. Serial dilutions from collected sample are made for OD600 measurement. Figure 5-5 presents the standard curve of C. jejuni growing on BHI broth plus 5 % NBS mixing with vancomycin and trimethoprim. While the standard curve of C. jejuni cultured on Mueller-Hinton broth with vancomycin and trimethoprim is shown in Figure 5-6.  83  Figure 5-5: Calibration curve of C. jejuni growing in BHI plus 5% NBS broth with vancomycin and trimethoprim. (Error bars were presented but too small to perceive).  84  Figure 5-6: Calibration curve of C jejuni growing in Mueller-Hinton broth with vancomycin and trimethoprim. (Error bars were presented but too small to perceive).  5.1.4 Mycobacterium smegmatis Similar to Campylobacter, M. smemgatis divides about every 3 hours. Thus, 18-hour was designed to achieve a logarithmic phase culture for build up the standard curve of M. smegmatis [59, 62].  85  The collected cells were washed twice with 10 mM PBS following with serial dilutions. Plate counting data was obtained after 2 days and was analyzed with corresponding absorbance values. Figure 5-7 shows a relationship between concentration (CFU/ml) and values of OD600 for M. smegmatis.  Figure 5-7: Calibration curve of M. smegmatis growing in Middlebrook 7H9 Broth. (Error bars were presented but too small to perceive).  In conclusion, comparing to S. typhimurium, E. coli enter into the exponential phase earlier and propagate faster ending up with a higher concentration after 2.5 hours under the  86  same culture conditions (215 rpm and 37 ℃). C. jejuni was cultured in different media to find out a better condition for culturing. Under the microaerophilic culture condition, cells grew better and had a higher concentration in BHI plus 5% NBS media than growing in Mueller-Hinton media. The reason can be due to the 5% NBS which is very nutrient ingredient for culturing C. jejuni. Based on the results showing above, BHI plus 5% NBS medium is preferred for growing C. jejuni. As aerobic bacteria, M. smegmatis grew even a little bit slower than C. jejuni but required less complex culture conditions.  5.2 Viability of Bacterial Cells in Diluted Buffer Low conductivities of suspending media are essential for generating p-DEP for trapping processes. In order to be captured, bacterial cells must experience surrounding buffers whose conductivities are much lower than those of the bacteria themselves [24]. Therefore, assessment of the viabilities of tested cells in diluted media fluid is an important index to determine the experimental time for trapping specific bacterial cells. This will prevent or decrease possibilities of misevaluating the trapping efficiency, which is of concern not only for live cells but dead ones as well. Two buffers — diluted PBS and diluted HEPES, are used as suspending buffer for testing. The conductivity of 10 mM PBS is about 1.8 mS/cm and the one of 10 mM HEPES is 0.98 mS/cm. In this experiment, the conductivities of 40 μS/cm are used to test the  87  survivability rate of specific bacterial cells, thus the concentrations of diluted buffers of PBS and HEPES are approximately 0.2 mM and 0.4 mM, respectively. Furthermore, the assessment of viability of bacterial cells in diluted buffer was only conducted using 40 μS/cm buffer but not 120 μS/cm. Because the higher conductivity buffer has higher concentration and 120 μS/cm is between the undiluted buffer (1.8 mS/cm or 0.98 mS/cm) and the 40 μS/cm buffer, it is expected that the results from the experiments using 40 μS/cm buffer will represent the viability assessment of the cells in the diluted buffer.  5.2.1 Salmonella typhimurium in Diluted PBS and HEPES Logarithmic-phase S. typhimurium cells were collected and prepared as described in Chapter 4. Cells were finally placed in microcentrifuge tubes filled with 0.2 mM PBS or 0.4 mM HEPES for viability testing and 10 mM PBS or HEPES for reference control. Every 5 minutes, 20 μl suspensions were sampled and transferred onto media plates after 10-fold serial dilutions. The plate-counting data were interpreted into percentage viability according to indicated time. Figure 5-8 and Figure 5-9 show the curves of survival rate (%) of S. typhimurium in diluted PBS and HEPES, respectively.  88  Figure 5-8: Viability of S. typhimurium in 0.2 mM PBS plus 10% dextrose.  89  Figure 5-9: Viability of S. typhimurium in 0.4 mM HEPES plus 10% dextrose.  The results demonstrate that regardless of what diluted buffer used for suspending S. typhimurium, the survival rate of the cells drop dramatically to about 30% for diluted PBS and about 15% for HEPES. Therefore, neither buffer is suitable for application as a suspending medium for trapping S. typhimurium. Further, 10% lost viability even appears in 10 mM PBS, which is supposed to maintain the viability of the cells for longer than 25 minutes.  90  5.2.2 Escherichia coli in Diluted PBS and HEPES E. coli growing at exponential phase were collected and prepared with the same procedures as those for S. typhimurium. The curve of survival rate of E. coli in diluted PBS is plotted in Figure 5-10.  Figure 5-10: Viability of E. coli in 0.2 mM PBS plus 10% dextrose.  From Figure 5-10, the time E. coli cells are able to endure in 0.2 mM PBS is limited — the cells start to loss viability as soon as the experiment begins, which means diluted PBS is not an appropriate buffer for evaluating the trapping efficiency of E. coli. In addition, the reference control of placing E. coli into 10 mM PBS seems to have an increasing trend,  91  which may be due to the random errors or a hypothesis that the cells proliferate slowly in PBS containing dextrose. On the other hand, when conducting the experiment by using diluted HEPES, the results demonstrate a large difference, as shown in Figure 5-11.  Figure 5-11: Viability of E. coli in 0.4 mM HEPES plus 10% dextrose.  The cells maintain relatively approximate survival rates for 15 minutes when placed in 0.4 mM HEPES, then drop to 30% viability as time passes. Compared to 0.2 mM PBS, diluted HEPES is more appropriate for application as a suspending medium for trapping experiments lasting less than 15 minutes. A similar escalating trend of viability from placing the cells in 10 mM PBS exists in 10 mM HEPES as well.  92  5.2.3 Campylobacter jejuni in Diluted PBS The survival rate of logarithmic-phase C. jujeni cells was tested in 10 mM and 0.2 mM PBS buffer for 25 minutes. The results are shown in Figure 5-12.  Figure 5-12: Viability of C. jejuni in 0.2 mM PBS plus 10% dextrose.  The viability almost drops closely to zero, which proves that the concentration of 0.2 mM PBS is too low, such that C. jejuni cells may lyse progressively as time passes. Based on the data, the cell membrane of E. coli is more vulnerable to that of either S. typhimurium or C. jejuni when subjected to diluted PBS.  93  5.2.4 Mycobacterium smegmatis in Diluted HEPES M. smegmatis cells of exponential phase were only tested in 0.4 mM HEPES. In Figure 5-13, the plot of the survival rate of cells against the indicated time is presented. Comparing to all the bacterial species examined within 0.2 mM HEPES, M. smegmatis shows a supreme performance of staying alive for at least 25 minutes in diluted buffer. This is very helpful for the experimental design of assessment of trapping efficiency.  Figure 5-13: Viability of M. smegmatis in 0.2 mM HEPES plus 10% dextrose.  94  5.3 DEP Capturing in Diluted PBS  5.3.1 Microspheres Capturing Since microspheres with fluorescent dye do not have any problem with losing viability in diluted PBS, experimental tests of assessing the functionality of the microconcentrator for trapping small particles with dimensions of about 2 μm was conducted with fluorescent beads suspended in PBS with conductivities of 20 μS/cm and 40 μS/cm, respectively. The beads were prepared and trapped in both 20 μS/cm and 40 μS/cm buffer, with better trapping appearing in the buffer with higher conductivity. At the conditions of 15 V peak- to- peak (pp) and 2 kHz, the beads started to be trapped in the diluted buffer of 20 μS/cm. But more beads suspended in 40 μS/cm buffer tended to be trapped by the electric field under the applied signal of 15 Vp-p and 2 kHz. As the frequency increased to 4 kHz, the beads began to come off from the electrodes. Flow rate was another important factor for trapping. Lower flow rate leads to a higher trapping efficiency. Figure 5-14 shows the beads’ trapping conditions with two different flow rates and frequencies both under 20 Vp-p.  95  Figure 5-14: Microspheres trapping in 0.2 mM PBS plus 10% dextrose. A. Trapping under 20 Vp-p and 1kHz with a flow rate of 100 μL/hr; B. 20 Vp-p and 1kHz with a flow rate of 500 μL/hr; 20 Vp-p and 3kHz with a flow rate of 100 μL/hr; 20 Vp-p and 3kHz with a flow rate of 500 μL/hr.  96  Under the same voltage, more beads were trapped with 1 kHz and slow flow rate (100 μL/hr), and as the flow rate increased up to 500 μL/hr, most of the beads were washed off from the electrodes even under lower frequency (1 kHz).  5.3.2 Salmonella typhimurium Capturing Additionally, S. typhimurium and C. jejuni were also tested in this microconcentrator to evaluate the capability of trapping. Trapping images are presented to prove that both S. typhimurium and C. jejuni can be trapped under certain conditions. For S. typhimurium, a suspension of ~107 CFU/mL with a conductivity of 30 μS/cm was introduced into the microconcentrator at a flow rate of 100 μL/hr. Different voltages (5 and 10 Vp-p) were applied with the same frequency (3MHz). The duration was controlled within 3 minutes. Trapping images are shown in Figure 5-15.  97  Figure 5-15: S. typhimurium trapping in 0.17 mM PBS plus 10% dextrose. A. Trapping under 5 Vp-p and 3 MHz with a flow rate of 100 μL/hr for 3 minutes; B. Trapping under 10 Vp-p and 3 MHz with a flow rate of 100 μL/hr for 3 minutes.  More S. typhimurium were trapped under larger voltage because the DEP force is directly proportional to the applied electric field according to Equation 2-1.  5.3.3 Campylobacter jejuni Capturing C. jejuni was tested in diluted PBS with different conductivities (30 μS/cm, 40 μS/cm and 50 μS/cm) under the same trapping conditions, as shown in Figure 5-16.  98  Figure 5-16: C. jejuni trapping in diluted PBS with different conductivities plus 10% dextrose. A. Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.17 mM buffer; B. Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.2 mM buffer; Trapping under 15 Vp-p and 500 kHz with a flow rate of 100 μL/hr for 3 minutes within 0.28 mM buffer.  The conductivity of the suspending buffer plays an important role in adjusting dielectrophoretic force. Using C. jejuni as an example, in Figure 5-16, the cells were preferentially captured in buffers with higher conductivity (50 μS/cm). However, comparing Figure 5-16A with Figure 5-16B, a higher conductivity buffer (40 μS/cm) did not provide a stronger force so as to provide a higher trapping efficiency, while a lower one (40 μS/cm) did.  99  5.4 Capturing Efficiency Evaluation The evaluations of trapping efficiency were only performed with E. coli and M. smegmatis but not with S. typhimurium and C. jejuni due to the results from the experiments of survival rate assessment. In addition, considering the complexity of culturing C. jejuni, the survival rate in diluted HEPES was not conducted as well as the assessment of trapping efficiency. Trapping efficiencies for both E. coli and M. smegmatis were calculated from the results of (input-output) / input concentration for serial values of applied frequencies under the same applied voltage (15 Vp-p) and flow rate. The conductivities of diluted HEPES buffer employed for acquiring the trapping efficiency were 40 μS/cm and 120 μS/cm.  5.4.1 Escherichia coli Capturing Based on the results of viability of E. coli in 0.4 mM HEPES (40 μS/cm), the experimental duration was controlled to 15 minutes including the washing process of the samples. Instead of using the same sample continuously for the whole set of experiments, a new sample was used for each applied frequency point. Figure 5-17 is a plot of trapping percentage (efficiency) versus applied frequency. As we can see, trapping percentage depends on applied frequency. In addition, trapping percentage also changes with different media conductivity. The lower conductivity curve  100  in black shows a decrease and then an increase in trapping percentage as the frequency is increased. The higher conductivity curve in red shows an increase, then decrease, and another increase in trapping percentage as the frequency is increased.  Figure 5-17: E. coli trapping efficiency in diluted HEPES of 40 μS/cm and 120 μS/cm plus 10% dextrose with a flow rate of 150 μL/hr.  Basically, for E. coli, the highest trapping efficiency with a higher applied frequency can be achieved in a lower conductivity (40 μS/cm) buffer, while within a higher conductivity (120 μS/cm) buffer, the highest capturing percentage can be obtained by  101  applying lower frequency. These two curves fluctuate symmetrically until they reach the frequency around 150 kHz, after which they have the same trend even with increasing frequency. This will be used as the applied frequency for estimating the effective conductivity of E. coli discussed in the next section.  5.4.2 Mycobacterium smegmatis Capturing Unlike E. coli, M. smegmatis is a Gram-positive bacterium and has a mycolic acid-constructed cell wall resulting in almost no losses in viability. In order to create comparable trapping efficiency results to those obtained for E. coli, the same trapping conditions (15Vp-p, 150 μL/hr and 10 minutes) were exploited for trapping M. smegmatis in diluted HEPES buffer of both 40 μS/cm and 120 μS/cm. Similarly, Figure 5-18 shows a plot of the trapping percentage (efficiency) versus applied frequency for M. smegmatis. As it is presented, in 40 μS/cm curve in black, trapping percentage remains relatively the same before 100 kHz and then increases rapidly after frequency increases. In the higher conductivity curve in red, the trapping percentage is high at first and then decreases, and finally increases with the same trend as the black curve.  102  Figure 5-18: M. smegmatis trapping efficiency in diluted HEPES of 40 μS/cm and 120 μS/cm plus 10% dextrose with a flow rate of 150 μL/hr.  The highest trapping percentage occurs at higher applied frequency in higher conductivity (120 μS/cm) buffer. The curves also achieve the same trend after certain frequency point, which is approximately over 900 kHz. In order to make sure that the frequency point used for measuring the effective conductivity of M. smegmatis has an equivalent trend after 900 kHz, 950 kHz was finally selected for measuring the effective conductivity of M. smegmatis.  103  5.5 Effective Conductivities Estimation Theoretically speaking, when the effective conductivity of a certain particle exceeds that of the suspending medium, p-DEP will become dominant. While it is lower than that of the suspending medium, the particle will direct to the weakest region of electric field, undergoing n-DEP. Measuring the effective conductivity of particles of interest is critical since particle segregation can be achieved by generating p-DEP for trapping the cells of interest from the mixture and producing n-DEP for repelling the remaining cells from the device. The effective conductivity of particle can be determined by using a simple optical measurement {Price, 1988 #46;J P H Burt , 1989 #45}. A relationship between the decreasing rates of optical absorbance ΔA and effective conductivity of particular particle can be written as [33],  ∆𝐴 = 𝑘  𝜎𝑝 −𝜎𝑚 𝜎𝑝 +2𝜎𝑚  (5-1)  where σp and σm are the conductivities of particle and medium respectively, k is a constant is a constant for a specific experiment and can be acquired by extrapolating σm to zero. A plot of ΔA versus a range of σm has to be plotted to determine σp. Figure 5-18 and  104  Figure 5-20 show the plots of the values of ΔA as a function of different values of σm for E. coli and M .smegmatis, respectively.  Figure 5-19: Values of absorbance change of E. coli as a function of the conductivities’ values of suspending media.  105  Figure 5-20: Values of absorbance change of M. smegmatis as a function of the conductivity values of the suspending media.  From Figure 5-19 and Figure 5-20, the value of k can be extrapolated to be 2.55 for E. coli and 2.24 for M. smegmatis when assuming σm to be 0.1 μS/cm. In addition, Equation 5-1 is rearranged to be,  106  𝜎𝑚 =  𝑘−∆𝐴 2∆𝐴+𝑘  𝜎𝑝  Then σp can be calculated out from the slope of a plot of σm versus  (5-2)  𝑘−∆𝐴 2∆𝐴+𝑘  .  Figure 5-21: Values of the suspending media conductivities as a function of (k - ΔA) / (2ΔA+k) for E. coli.  107  Figure 5-22: Values of the suspending media conductivities as a function of (k - ΔA) / (2ΔA+k) for M. smegmatis.  From Figure 5-21, the value of E. coli conductivity is approximately assessed is calculated as 610.7 μS/cm, and the conductivity value of M. smegmatis was derived from Figure 5-22, and equals 1095 μS/cm. There are several variabilities existing in this method, leading to variable results. It is very hard to determine the k value since σm will never be zero from the graph of ΔA versus σm. The k value can be only estimated by extrapolation to  108  zero, the smaller the k value is, the bigger the σp can be obtained. That is not a practically accurate method to determine the effective conductivity of a specific microorganism. The conductivity value of E. coli calculated from my experiments is much higher than the one of E. coli ATCC 11303 calculated by Markx et al., which is 412±25μS/cm [33]. This might be due to the difference of the strains used in the experiment or the difference of estimating the value of k. The conductivity of M. smegmatis is not comparable since there is not similar calculation from any literature so far.  5.6 Mixed Sample Partitioning According to the theory of separating microorganisms introduced by Markx et al.[33], and the results obtained from the calculation of effective conductivities of E. coli and M. smegmatis, separation of these two species from one another was attempted by using a suspending buffer of conductivity value between the effective conductivities of E. coli and M. smegmatis, which is about 700 μS/cm and an applied frequency at the middle of the those applied to determine the effective conductivities of E. coli and M. smegmatis, which is about 450 kHz. The experiment was conducted using the two-set-electrode chip. The first electrode was powered at 8Vp-p and 450 kHz while the second one was powered at with 8 Vp-p and 900 kHz. The HEPES buffer plus 10% dextrose with conductivity value of 700 μS/cm was  109  used as suspending medium. E. coli and M. smegmatis with the same concentration were first injected into the device for 40 minutes individually. As a very high conductivity HEPES buffer was used as suspending medium, it is believed that the viability of E. coli will be better than that observed under lower conductivity buffer conditions of 40 μS/cm. Thus, an assumption was made that viability of E. coli would be still maintained at the original level after 40 minutes. Next, a mixed sample of E. coli and M. smegmatis was prepared with 1:1 concentration and flowed into the single channel with two-set-electrode chip with a flow rate of 150 μL/hr for 40 minutes. The separation results were record by imaging through the CCD camera. Figure 5-23 and Figure 5-24 show the images of two individual experiments operating on two sets of electrodes. E. coli cells (Figure 5-23A and C, Figure 5-24A and C) were trapped with an applied voltage of 8 Vp-p and frequency of both 450 kHz and 900 kHz after 40 minutes while there were almost no M. smegmatis cells (Figure 5-23B and D, Figure 5-24B and D) were observed to be trapped on the electrodes.  110  Figure 5-23: E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose individually. A. E. coli trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B.  111  Figure 5-24: E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose individually. A. E. coli trapped on the second set of electrodes at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the second set of electrodes at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B.  On the other hand, mixture of E. coli and M. smegmatis with the concentration of 1:1 was introduced into the channel with 8 Vp-p at 450 kHz for the set of electrode and 900 kHz for the second one. The observing results are presented in Figure 5-25 and Figure 5-26. Both sets of electrodes trapped E. coli cells down by p-DEP force and the second set applied at 900 kHz captured even more cells than the first set with lower applied frequency.  112  In terms of M. smegmatis trapping conditions, there were some blurry bands on the first set of electrode (Figure 5-25B and D) which are believed to be the non-specific trapping or auto-fluorescent impurities located in the PDMS or on the substrate but not GFP-expressing M. smegmatis cells. In Figure 5-26B and D, some spots could be observed sitting on the second set of electrode, which are not M. smegmatis cells either, but bubbles due to the appearance of electrolysis.  113  Figure 5-25: Mixture of E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose together. A. E. coli trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μl/hr; B. M. smegmatis trapped on the first set of electrode at 8 Vp-p and 450 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B.  114  Figure 5-26: Mixture of E. coli and M. smegmatis trapping in diluted HEPES of 700 μS/cm plus 10% dextrose together. A. E. coli trapped on the second set of electrode at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; B. M. smegmatis trapped on the second set of electrode at 8 Vp-p and 900 kHz with a flow rate of 150 μL/hr; C. Contrast color of A; D. Contrast color of B.  According to the effective conductivities calculated from the last section, it was initially assumed that the E. coli cells would not be trapped but the M. smegmatis cells would. The effective conductivity of E. coli DH5α was calaculated to be only 610.7 μS/cm, which should be less than that of the suspending medium. Moreover, M. smegmatis cells  115  have a much higher effective conductivity value than E. coli cells as well as that of the applied suspending medium, such that M. smegmatis cells were supposed to be trapped. However, the results obtained from the experiments suggest the opposite of the original hypothesis of separation. More research on these results is clearly indicated. It is most likely that the methods used for calculating the conductivity of the cells in Section 5.5 was flawed due to the inaccuracies noted. A more robust method for measuring effective conductivity may be electrorotation, which has been used for such purposes in the past. Nevertheless, we have demonstrated that it is possible to screen out one population of cells from another using our microfluidic DEP system. Further manipulation of the operating conditions will hopefully reverse the trapping such that we are able to trap pathogenic M smegmatis and not E. coli.  5.7 DEP Capturing in Bodily Fluid Pooled pathogen-negative CSF was used as the suspending buffer for trapping both E. coli and M. smegmatis through simple dilution to achieve certain conductivities for trapping. In order to acquire the same concentration of bacterial cells traveling through the channel with different flow rates, the experimental durations were calibrated correspondingly. Images were taken with 5X and 10X magnifications and are shown below  116  Figure 5-27: E. coli trapping in diluted CSF at 120 μS/cm. A. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 5X magnification; B. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 5X magnification; C. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 5X magnification; D. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150μL/hr for 300 seconds with 10X magnification; E. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 10X magnification; F. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 10X magnification.  117  Figure 5-28: M. smegmatis trapping in diluted CSF of 120 μS/cm. A. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 5X magnification; B. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 5X magnification; C. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 5X magnification; D. Trapping under 15 Vp-p and 1 MHz with a flow rate of 150 μL/hr for 300 seconds with 10X magnification; E. Trapping under 15 Vp-p and 1 MHz with a flow rate of 750 μL/hr for 60 seconds with 10X magnification; F. Trapping under 15 Vp-p and 1 MHz with a flow rate of 1500 μL/hr for 30 seconds with 10X magnification.  118  In Figure 5-27, the largest number of E. coli cells are trapped with the lowest flow rate of 150 μL/hr. Fewer cells are trapped as the flow rate increases. Similarly, exponential-phase  M.  smegmatis  cells  are  prepared  and  injected  into  the  microconcentractor with different flow rates for corresponding duration. Images were taken as shown in Figure 5-28. Again, more M. smegmatis cells were trapped under a slow flow rate while fewer were trapped as the flow rate increased. This is once more showing evidence that the flow rate is an essential factor affecting the trapping efficiency. Generally speaking, the lower flow rate applied, the higher trapping percentage can be obtained. Future research will apply conclusions obtained from effective conductivity measurements to design operating conditions such that M. smegmatis alone will be trapped from CSF, resulting in a rapid, low-power diagnostic system for mycobacterial meningitis.  119  CHAPTER 6: CONCLUSIONS 6.1 Summary The interdigitated type of electrodes was chosen because of its simplicity and popularity of fabrication and application. The positive dielectrophretic force generated from the microelectrodes has to be strong enough to break the force balance in order to hold the particles from dragging away by the flow force or floating upward by the buoyancy force. In addition, p-DEP force needs to be created under the environment of lower conductivity of the suspending medium than that of the particles. Therefore, the impact of lower conductivity buffer to testing particles was essential to evaluate the functionality and performance of this microdevice. The results showed that diluted PBS as suspending buffer has strong effects on the viability of bacterial cells that more than 60% of them would die within 25 minutes, while diluted HEPES buffer has a better performance on maintaining the viability of the cells. Characterizations of trapping conditions for E. coli and M. smegmatis were achieved by assessing the trapping efficiency within a suspending medium of two different conductivities. The results from this experiment are also beneficial for segregating one of them from a mixture. Further, effective conductivities of E. coli and M. smegmatis were determined by using a method of measuring optical density [33]. From these data, an  120  appropriate conductivity of suspending media for partitioning the E. coli or M. smegmatis from a mixture of them was established. However, conflict between the experimental results and the hypothesis existed. The E. coli cells were trapped while the M. smegmatis cells were flushed through, which was opposite to the expected result. Further work on determining the effective conductivities of bacterial populations is required to effectively set operating conditions for trapping one desired species compared to others. In this work, the most challenging issue encountered was the problem of non-specific adhesion of unknown particles to the surface of the bottom in the channel. Bovine serum albumin (BSA) was applied the gold electrode surface and Tween® 20 was added into testing samples for eliminating or decreasing the non-specific adhesion. Nevertheless, after several operating experiments on the same microchip, the problem of non-specific adhesion started to appear and accumulation that intensely affect the accuracy of defining the trapping efficiency was observed. Some non-specific bonding on the PDMS channel was also a problem and could result in confusing the visual inspection. The other challenges encountered were the difficulties of getting access to the optical observation for certain species of bacteria due to their weak GFP-promoters, and determining some important information including effective conductivities, which are crucial for optimizing the separation process. Successful demonstration of operating this microconcentrator for trapping polystyrene microspheres, cells of S. typhimurium, E. coli, C. jejuni and M. smegmatis was performed.  121  These results will assist in building up a model for concentrating and purifying the targeted bacteria of interest in more complicated samples including bodily fluids or stool samples. More improvements and optimizations have to be continued on to enhance the performance of this device.  6.2 Future Work Current trapping performance is limited by low throughput. An optimized chaotic mixer with staggered herringbone (Figure 6-1) [65] will be essential to achieving a high flow rate and better trapping efficiency by dramatically increasing the mass transport of the testing particles to the electrodes.  122  Figure 6-1: ―Device overview. (A) Illustration of the assembled devices, showing the PDMS channel and gold IDEs on the Pyrex substrate. (B) Illustration of particle trapping in laminar flow and chaotic flow. Without the mixer, only a portion of particles carried by the laminar flow are exposed to DEP and trapped by electrodes. Chaotic flow generated by a passive micromixer circulates the particles and exposes more particles to the DEP region. (C) Schematic of the three kinds of micromixer geometries described in this work. From left to right: slanted groove micromixer (SGM), herringbone micromixer (HM), and staggered herringbone micromixer (SHM).‖ [65]  Based on results acquired from Lee and Voldman’s study [65], this staggered herringbone mixer can achieve about 1.5 X of trapping efficiency comparing to a smooth channel. In addition, a more robust and reusable design will help to prevent any leaks between the PDMS and the substrate or any non-specific adhesions on the PDMS. Glass is preferred to fabricate the whole microconcentrator due to its durability and biocompatibility. A new design of the PCB working as the interface between the power source and the micro device must be fabricated in order to achieve the simplicity of aligning and mounting  123  the microdevice on board. Multiple channels with valves control are recommended for increasing the testing efficiency and enriching the experimental designs. In the future, the ultimate goal will be constructing an all-in-one DEP microsystem incorporating downstream genetic detection after the DEP concentration, purification and differentiation with advantages of portability, low cost, low power, and high-throughput in a rapid detection manner. Such a diagnostic microsystem is desired for detecting rare infectious pathogens such as M. tuberculosis from crude samples and more importantly, creating an important interface between the pathogenic analysis and the timely clinical health intervention or appropriate patient treatments when pathogenic outbreaks occur.  124  BIBLIOGRAPHY [1]  P. B. Eckburg, E. M. Bik, C. N. Bernstein et al., ―Diversity of the Human Intestinal Microbial Flora,‖ Science, vol. 308, no. 5728, pp. 1635-1638, June 10, 2005, 2005.  [2]  K. D. Dunkley, T. R. Callaway, V. I. Chalova et al., ―Foodborne Salmonella ecology in the avian gastrointestinal tract,‖ Anaerobe, vol. 15, no. 1-2, pp. 26-35, 2009.  [3]  T. Slanec, A. Fruth, K. Creuzburg et al., ―Molecular Analysis of Virulence Profiles and Shiga Toxin Genes in Food-Borne Shiga Toxin-Producing Escherichia coli,‖ Appl. Environ. Microbiol., vol. 75, no. 19, pp. 6187-6197, October 1, 2009, 2009.  [4]  L. H. Gould, L. Demma, T. F. Jones et al., ―Hemolytic Uremic Syndrome and Death in Persons with Escherichia coli O157:H7 Infection, Foodborne Diseases Active Surveillance Network Sites, 2000-2006,‖ Clinical Infectious Diseases, vol. 49, no. 10, pp. 1480-1485, 2009.  [5]  T. Duong, and M. E. Konkel, ―Comparative studies of Campylobacter jejuni genomic diversity reveal the importance of core and dispensable genes in the biology of this enigmatic food-borne pathogen,‖ Current Opinion in Biotechnology, vol. 20, no. 2, pp. 158-165, 2009.  [6]  S. Péhin, C. Janoir, H. Boureau et al., ―Diminished intestinal colonization by Clostridium difficile and immune response in mice after mucosal immunization  125  with surface proteins of Clostridium difficile,‖ Vaccine, vol. 25, no. 20, pp. 3946-3954, 2007. [7]  H. Koo, N. Ajami, Z. Jiang et al., ―A Nosocomial Outbreak of Norovirus Infection Masquerading as Clostridium difficile Infection,‖ Clinical Infectious Diseases, vol. 48, no. 7, pp. e75-e77, 2009.  [8]  CBCNews,  ―Montreal  hospital  battles  superbug,‖  http://www.cbc.ca/canada/montreal/story/2010/03/01/montreal-c-difficile-outbrea k-rosemont-hospital.html?ref=rss, Monday, March 1, 2010. [9]  L. G. Miller, F. Perdreau-Remington, A. S. Bayer et al., ―Clinical and Epidemiologic  Characteristics  Methicillin-Resistant  Cannot  Staphylococcus  Distinguish  Community-Associated  aureus  Infection  from  Methicillin-Susceptible S. aureus Infection: A Prospective Investigation,‖ Clinical Infectious Diseases, vol. 44, no. 4, pp. 471-482, 2007. [10]  P. Kara, B. Meric, and M. Ozsoz, ―Application of Impedimetric and Voltammetric Genosensor for Detection of a Biological Warfare: Anthrax,‖ Electroanalysis, vol. 20, no. 24, pp. 2629-2634, 2008.  [11]  WorldHealthOrganization, ―Global tuberculosis control: a short update to the 2009 report,‖ 2009.  [12]  B. W. Brooks, J. Devenish, C. L. Lutze-Wallace et al., ―Evaluation of a monoclonal antibody-based  enzyme-linked  immunosorbent  126  assay  for  detection  of  Campylobacter fetus in bovine preputial washing and vaginal mucus samples,‖ Veterinary Microbiology, vol. 103, no. 1-2, pp. 77-84, 2004. [13]  P. M. Fratamico, ―Comparison of culture, polymerase chain reaction (PCR), TaqMan Salmonella, and Transia Card Salmonella assays for detection of Salmonella spp. in naturally-contaminated ground chicken, ground turkey, and ground beef,‖ Molecular and Cellular Probes, vol. 17, no. 5, pp. 215-221, 2003.  [14]  F. O. Finlay, H. Witherow, and P. T. Rudd, ―Latex agglutination testing in bacterial meningitis,‖ Archives of Disease in Childhood, vol. 73, no. 2, pp. 160-161, August 1995, 1995.  [15]  K. NM., ―The value of latex particle agglutination test for rapid detection of bacterial antigens in the cerebrospinal fluid.,‖ Saudi Med J. , vol. 29, no. 5, pp. 774-5., May, 2008.  [16]  J. R. Crowther, ―ELISA. Theory and practice.,‖ Methods Mol Biol., vol. 42, pp. 1-218, 1995.  [17]  O. Lazcka, F. J. D. Campo, and F. X. Muñoz, ―Pathogen detection: A perspective of traditional methods and biosensors,‖ Biosensors and Bioelectronics, vol. 22, no. 7, pp. 1205-1217, 2007.  [18]  M. D. Zordan, M. M. G. Grafton, G. Acharya et al., ―Detection of pathogenic E. coli O157:H7 by a hybrid microfluidic SPR and molecular imaging cytometry device,‖ Cytometry Part A, vol. 75A, no. 2, pp. 155-162, 2009.  127  [19]  N. Beyor, T. Seo, P. Liu et al., ―Immunomagnetic bead-based cell concentration microdevice for dilute pathogen detection,‖ Biomedical Microdevices, vol. 10, no. 6, pp. 909-917, 2008.  [20]  H.-C. Chang, ―Nanobead electrokinetics: The enabling microfluidic platform for rapid multi-target pathogen detection,‖ AIChE Journal, vol. 53, no. 10, pp. 2486-2492, 2007.  [21]  H. A. Pohl, ―The Motion and Precipitation of Suspensions in Divergent Electric Field,‖ Journal of Applied Physics vol. 22 no. 7, pp. 869 - 871 Jul 1951.  [22]  H. A. Pohl, ―Dielectrophoresis the behavior of neutral matter in nonuniform electric fields,‖ Cambridge University Press, 1978.  [23]  T. B. Jones, ―Electromechanics of Particles,‖ Cambridge University Press, 1995.  [24]  J. Voldman, "Dielectrophoretic Traps for Cell Manipulation," BioMEMS and Biomedical Nanotechnology, pp. 159-186, 2007.  [25]  B. Crowell, ―Simple Nature,‖ Light and Matter, 2008.  [26]  H. J. Curtis, and K. S. Cole, ―TRANSVERSE ELECTRIC IMPEDANCE OF THE SQUID GIANT AXON,‖ The Journal of General Physiology, vol. 21, no. 6, pp. 757-765, July 20, 1938, 1938.  [27]  R. Pethig, and G. H. Markx, ―Applications of dielectrophoresis in biotechnology,‖ Trends in Biotechnology, vol. 15, no. 10, pp. 426-432, 1997.  128  [28]  A. Irimajiri, T. Hanai, and A. Inouye, ―A dielectric theory of "multi-stratified shell" model with its application to a lymphoma cell,‖ Journal of Theoretical Biology, vol. 78, no. 2, pp. 251-269, 1979.  [29]  P. R. C. Gascoyne, F. F. Becker, and X. B. Wang, ―Numerical analysis of the influence of experimental conditions on the accuracy of dielectric parameters derived  from  electrorotation  measurements,‖  Bioelectrochemistry  and  Bioenergetics, vol. 36, no. 2, pp. 115-125, 1995. [30]  Peter R. C. Gascoyne and Jody V. Vykoukal, ―Dielectrophoresis-Based Sample Handling in General-Purpose Programmable Diagnostic Instruments,‖ Proc IEEE Inst Electr Electron Eng., vol. 92(1), pp. 22-42, January 1, 2004.  [31]  H. Gram, ―Über die isolierte Färbung der Schizomyceten in Schnitt- und Trockenpräparaten,‖ Fortschritte der Medizin, vol. 2, pp. 185-9, 1884.  [32]  L.-T. Ou, and R. E. Marquis, ―Electromechanical Interactions in Cell Walls of Gram-Positive Cocci,‖ J. Bacteriol., vol. 101, no. 1, pp. 92-101, January 1, 1970, 1970.  [33]  G. H. Marks, Y. Huang, X.-F. Zhou et al., ―Dielectrophoretic characterization and separation of micro-organisms,‖ Microbiology, vol. 140, no. 3, pp. 585-591, March 1, 1994.  129  [34]  I. Doh, and Y.-H. Cho, ―A continuous cell separation chip using hydrodynamic dielectrophoresis (DEP) process,‖ Sensors and Actuators A: Physical, vol. 121, no. 1, pp. 59-65, 2005.  [35]  E. T. Lagally, S.-H. Lee, and H. T. Soh, ―Integrated microsystem for dielectrophoretic cell concentration and genetic detection,‖ Lab on a Chip, vol. 5, no. 10, pp. 1053-1058, 2005.  [36]  N. Gadish, and J. Voldman, ―High-Throughput Positive-Dielectrophoretic Bioparticle Microconcentrator,‖ Analytical Chemistry, vol. 78, no. 22, pp. 7870-7876, 2006.  [37]  C. M. Das, F. Becker, S. Vernon et al., ―Dielectrophoretic Segregation of Different Human Cell Types on Microscope Slides,‖ Analytical Chemistry, vol. 77, no. 9, pp. 2708-2719, 2005.  [38]  F. Grom, J. Kentsch, T. Müler et al., ―Accumulation and trapping of hepatitis A virus  particles  by  electrohydrodynamic  flow  and  dielectrophoresis,‖  ELECTROPHORESIS, vol. 27, no. 7, pp. 1386-1393, 2006. [39]  M. S. Jaeger, and et al., ―Contact-free single-cell cultivation by negative dielectrophoresis,‖ Journal of Physics D: Applied Physics, vol. 41, no. 17, pp. 175502, 2008.  130  [40]  L. Ying, S. S. White, A. Bruckbauer et al., ―Frequency and Voltage Dependence of the Dielectrophoretic Trapping of Short Lengths of DNA and dCTP in a Nanopipette,‖ Biophysical Journal, vol. 86, no. 2, pp. 1018-1027, 2004.  [41]  U. Kim, J. Qian, S. A. Kenrick et al., ―Multitarget Dielectrophoresis Activated Cell Sorter,‖ Analytical Chemistry, vol. 80, no. 22, pp. 8656-8661, 2008.  [42]  M. Castellarnau, A. Errachid, C. Madrid et al., ―Dielectrophoresis as a Tool to Characterize and Differentiate Isogenic Mutants of Escherichia coli,‖ Biophysical Journal, vol. 91, no. 10, pp. 3937-3945, 2006.  [43]  S. Franssila, ―Introduction to Microfabrication,‖ May 2004.  [44]  M. J. Madou, ―Fundamentals of Microfabrication: The Science of Miniaturization, Second Edition,‖ Taylor & Francis, Inc., March, 2002.  [45]  M. Meyyappan, and et al., ―Carbon nanotube growth by PECVD: a review,‖ Plasma Sources Science and Technology, vol. 12, no. 2, pp. 205, 2003.  [46]  A. D. S. Bhattacharya, J.M. Berg, S. Gangopadhyay, ―Studies on surface wettability of puly(dimethyl) siloxane (PDMS) and glass under oxygen plasma treatment and correlation with bond strength,‖ Journal of Micro Electro Mechanical Systems, vol. 14(3), pp. 590-597, 2005.  [47]  H. Hillborg, and U. W. Gedde, ―Hydrophobicity changes in silicone rubbers,‖ IEEE Transactions on Dielectrics and Electrical Insulation, vol. 6, pp. 703 - 717, 1999.  131  [48]  C. L. Mirley, and J. T. Koberstein, ―A Room Temperature Method for the Preparation of Ultrathin SiOx Films from Langmuir-Blodgett Layers,‖ Langmuir, vol. 11, no. 4, pp. 1049-1052, 1995.  [49]  Y. Berdichevsky, J. Khandurina, A. Guttman et al., ―UV/ozone modification of poly(dimethylsiloxane) microfluidic channels,‖ Sensors and Actuators B: Chemical, vol. 97, no. 2-3, pp. 402-408, 2004.  [50]  S. C. Baicu, and M. J. Taylor, ―Acid-base buffering in organ preservation solutions as a function of temperature: new parameters for comparing buffer capacity and efficiency,‖ Cryobiology, vol. 45, no. 1, pp. 33-48, 2002.  [51]  E. N. Marieb, and J. Mallatt, ―Human Anatomy, Second Edition,‖ Benjamin/Cummings Pub. Co, 1996.  [52]  B. Madison, ―Application of stains in clinical microbiology,‖ Biotechnic and Histochemistry, vol. 76, no. 3, pp. 119 - 125, 2001.  [53]  G. Etienne, F. Laval, C. Villeneuve et al., ―The cell envelope structure and properties of Mycobacterium smegmatis mc2155: is there a clue for the unique transformability of the strain?,‖ Microbiology, vol. 151, no. 6, pp. 2075-2086, June 1, 2005, 2005.  [54]  R. E. Gordon, and M. M. Smith, ―RAPIDLY GROWING, ACID FAST BACTERIA I.Species' Descriptions of Mycobacterium phlei Lehmann and  132  Neumann and Mycobacterium smegmatis (Trevisan) Lehmann and Neumann,‖ J Bacteriol., vol. 69(5), pp. 502–507, May, 1955. [55]  T. Hong, W. R. Butler, F. Hollis et al., ―Characterization of a Novel Rapidly Growing Mycobacterium Species Associated with Sepsis,‖ J. Clin. Microbiol., vol. 41, no. 12, pp. 5650-5653, December 1, 2003.  [56]  J. A. Newton, Jr., P. J. Weiss, W. A. Bowler et al., ―Soft-Tissue Infection Due to Mycobacterium smegmatis: Report of Two Cases,‖ Clinical Infectious Diseases, vol. 16, no. 4, pp. 531-533, 1993.  [57]  J. S. Tyagi, and D. Sharma, ―Mycobacterium smegmatis and tuberculosis,‖ Trends in Microbiology, vol. 10, no. 2, pp. 68-69, 2002.  [58]  L. J. Alderwick, H. L. Birch, A. K. Mishra et al., ―Structure, function and biosynthesis of the Mycobacterium tuberculosis cell wall: arabinogalactan and lipoarabinomannan assembly with a view to discovering new drug targets,‖ Biochemical Society Transactions, vol. 035, no. 5, pp. 1325-1328, October 1, 2007.  [59]  A. K. Singh, and J. M. Reyrat, ―Laboratory Maintenance of Mycobacterium smegmatis,‖ Current Protocols in Microbiology, vol. UNIT 10C.1 August, 2009.  [60]  M. E. Konkel, J. E. Christensen, A. S. Dhillon et al., ―Campylobacter jejuni Strains Compete for Colonization in Broiler Chicks,‖ Appl. Environ. Microbiol., vol. 73, no. 7, pp. 2297-2305, April 1, 2007.  133  [61]  B. Sampathkumar, S. Napper, C. D. Carrillo et al., ―Transcriptional and translational expression patterns associated with immobilized growth of Campylobacter jejuni,‖ Microbiology, vol. 152, no. 2, pp. 567-577, February 1, 2006.  [62]  R. Gadagkar, and K. Gopinathan, ―Growth ofMycobacterium smegmatis in minimal and complete media,‖ Journal of Biosciences, vol. 2, no. 4, pp. 337-348, 1980.  [63]  J. A. R. Price, J. P. H. Burt, and R. Pethig, ―Applications of a new optical technique for measuring the dielectrophoretic behaviour of micro-organisms,‖ Biochimica et Biophysica Acta (BBA) - General Subjects, vol. 964, no. 2, pp. 221-230, 1988.  [64]  J. P. H. Burt, and et al., ―An optical dielectrophoresis spectrometer for low-frequency measurements on colloidal suspensions,‖ Journal of Physics E: Scientific Instruments, vol. 22, no. 11, pp. 952, 1989.  [65]  H.-Y. Lee, and J. Voldman, ―Optimizing Micromixer Design for Enhancing Dielectrophoretic Microconcentrator Performance,‖ Analytical Chemistry, vol. 79, no. 5, pp. 1833-1839, 2007.  134  APPENDICES Appendix A. Schematic of SMT Board  135  Appendix B. Raw Data of Growth Curves and Calibration Curves S. typhimurium cultured in LB media (Sample Growth Curve). Time(min) 33 68 90 120 200  CFU/mL 1.50E+07 1.95E+07 2.20E+07 2.90E+07 1.30E+08  S. typhimurium cultured in LB media (Sample Calibration Curve). CFU/mL 1.50E+07 1.95E+07 2.20E+07 2.90E+07 1.30E+08  OD600 0.08925 0.1475 0.2745 0.57275 1.555  E. coli cultured in LB media (Sample Growth Curve). Time(min) 33 68 90 120 200  CFU/mL 1.80E+07 3.84E+07 8.25E+07 1.25E+08 2.2E+08  136  E. coli cultured in LB media (Calibration Curve). 1st CFU/mL 3.61E+07 6.02E+07 7.23E+07 9.03E+07  OD600 0.3255 0.5045 0.608  2nd CFU/ml 1.80E+07 3.84E+07 8.25E+07  OD600 0.1765 0.3245 0.5665  0.7695  1.25E+08  0.9705  137  C. jejuni cultured in BHI+5% NBS media (Sample Calibration Curve). OD600 Measurements 0.104 0.181 0.3 0.386 0.626  0.106 0.178 0.298 0.385 0.621  OD600 Average  0.107 0.179 0.299 0.385 0.62  0.105667 0.179333 0.299 0.385333 0.622333  CFU/mL 1.35E+07 2.43E+07 4.04E+07 4.85E+07 8.09E+07  C. jejuni cultured in MH media (Sample Calibration Curve). OD600 Measurements 0.05 0.074 0.116 0.15 0.188 0.313  0.047 0.075 0.116 0.15 0.189 0.31  OD600 Average 0.046 0.071 0.117 0.151 0.19 0.307  138  0.047667 0.073333 0.116333 0.150333 0.189 0.31  CFU/mL 1.26E+06 2.27E+06 2.84E+06 3.79E+06 4.55E+06 7.58E+06  M. smegmatis cultured in 7H9 media. OD600 Measurements 2.389 0.991 0.444 0.256 0.026  2.322 0.951 0.435 0.251 0.021  139  OD600 Average  CFU/mL  2.3555 0.971 0.4395 0.2535 0.0235  6.07E+08 3.03E+08 1.52E+08 6.07E+07 6.07E+06  Appendix B. Raw Data of Viability Assessment S. typhimurium in diluted PBS (0.2 mM and 40 μS/cm). Time(min)  10 mM PBS(CFU/mL)  0 5 10 15 20 25  2.750E+04  1st 2nd 0.2 mM PBS(CFU/mL) 0.2 mM PBS(CFU/mL) 2.250E+04 1.575E+04 1.250E+04 8.750E+03 7.250E+03 8.750E+03  2.500E+04  2.650E+04 1.800E+04 1.450E+04 1.075E+04 7.750E+03 7.750E+03  S. typhimurium in diluted HEPES (0.4mM and 40 μS/cm). st  st  nd  nd  Time (min)  1 10 mM HEPES (CFU/mL)  1 0.4 mM HEPES (CFU/mL)  2 10 mM HEPES (CFU/mL)  2 0.4 mM HEPES (CFU/mL)  0 5 10 15 20 25  1.77E+06  2.37E+05 1.25E+05 5.60E+04 3.90E+04 4.23E+04 2.97E+04  5.13E+05  2.53E+06 1.77E+06 5.47E+05 6.93E+05 2.77E+05 6.30E+05  1.63E+06  3.78E+05  E. coli in diluted PBS (0.2mM and 40 μS/cm). Time(min)  10mM PBS(CFU/mL)  0 5 10 15 20 25  3.925E+06  1st 2nd 0.2 mM PBS(CFU/mL) 0.2 mM PBS(CFU/mL) 4.750E+06 1.150E+06 3.775E+05 1.575E+05 1.000E+05 5.500E+04  4.213E+06  140  4.925E+06 1.200E+06 3.100E+05 1.450E+05 7.700E+04 4.750E+04  E. coli in diluted HEPES (0.4mM and 40 μS/cm). st  T(min) 0 5 10 15 20 25  st  nd  nd  1 10 mM HEPES (CFU/mL)  1 0.4 mM HEPES (CFU/mL)  2 10 mM HEPES (CFU/mL)  2 0.4 mM HEPES (CFU/mL)  4.73E+06  3.87E+06 4.67E+06 4.70E+06 4.40E+06 1.70E+06 1.60E+06  5.10E+06  3.00E+06  4.65E+06  T(min)  3.33E+06 2.53E+06 1.87E+06  6.04E+06  3rd 10 mM HEPES (CFU/mL)  3rd 0.4 mM HEPES (CFU/mL)  8.00E+04  9.67E+04  8.59E+04  9.00E+04 4.30E+04 2.00E+04  0 5 10 15 20 25  C. jejuni in diluted PBS (0.2mM and 40 μS/cm). T(min)  10mM PBS(CFU/mL)  0 5 10 15 20 25  1.025E+05  1st 2nd 0.2 mM PBS(CFU/mL) 0.2 mM PBS(CFU/mL) 1.250E+05 6.250E+04 1.925E+04 8.670E+03 5.500E+03 3.250E+02  1.225E+05  1.000E+05 4.000E+04 2.200E+04 8.500E+03 6.500E+03 6.750E+02  M. smegmatis in diluted HEPES (0.4mM and 40 μS/cm). T(min)  1st 10 mM HEPES (CFU/mL)  1st 0.4 mM HEPES (CFU/mL)  141  2nd 10 mM HEPES (CFU/mL)  2nd 0.4 mM HEPES (CFU/mL)  0 5 10 15 20 25  2.47E+06 2.67E+06 2.30E+06 2.37E+06 2.37E+06 2.37E+06  2.30E+06 2.40E+06 2.63E+06 2.80E+06 2.87E+06 2.60E+06  142  2.80E+06 1.95E+06 2.70E+06 2.55E+06 2.40E+06 2.30E+06  2.65E+06 2.40E+06 2.40E+06 2.40E+06 2.47E+06 2.40E+06  Appendix C. Raw  Data of  Trapping  Efficiency  Evaluation Trapping Percentage vs. applied frequency for E. coli in 40 μS/cm HEPES. Frequency(Hz)  1st Trapping Percentage  2nd Trapping Percentage  Average Trapping Percentage  3333 10000 33333 100000 333333 1000000  51.85% 59.10% 43.99% 67.27% 68.13% 74.44%  64.48% 60.44% 46.18% 70.24% 77.56% 89.47%  58.16% 59.77% 45.09% 68.76% 72.84% 81.96%  Trapping Percentage vs. applied frequency for E. coli in 120 μS/cm HEPES. Frequency(Hz)  1st Trapping Percentage  2nd Trapping Percentage  Average Trapping Percentage  3333 10000 33333 100000 333333 1000000  62.25% 58.34% 73.52% 43.59% 54.82% 57.38%  61.06% 50.36% 62.19% 31.18% 46.25% 61.34%  61.65% 54.35% 67.86% 37.38% 50.54% 59.36%  143  Trapping Percentage vs. applied frequency for M. smegmatis in 40 μS/cm HEPES. st  nd  Frequency(Hz)  1 Trapping Percentage  2 Trapping Percentage  Average Trapping Percentage  3333 10000 33333 100000 333333 1000000  21.76% 27.95% 29.02% 29.93% 41.61% 58.88%  26.25% 31.20% 30.19% 30.62% 49.69% 66.73%  24.00% 29.57% 29.60% 30.27% 45.65% 62.81%  Trapping Percentage vs. applied frequency for M. smegmatis in 120 μS/cm HEPES. Frequency(Hz)  1st Trapping Percentage  2nd Trapping Percentage  Average Trapping Percentage  3333 10000 33333 100000 333333 1000000  58.13% 17.76% 50.53% 58.26% 67.07% 70.08%  53.55% 20.22% 47.73% 55.74% 45.53% 71.03%  55.84% 18.99% 49.13% 57.00% 56.30% 70.56%  144  Appendix D. Raw Data of Effective Conductivities Estimation Absorbance change vs. various conductivities of suspending media for E. coli. Conductivity (μ S/cm)  Change in Absorbance (ΔA)  40 60 80 120 160 200 320  0.744056448 0.582026755 0.523238025 0.388192536 0.352545171 0.271346386 0.062760039  Various conductivities of suspending media vs. (k-ΔA)/ (2ΔA+k) for E. coli. k  Conductivity (μ S/cm)  (k-ΔA)/(2ΔA+k)  2.554078  40 60 80 120 160 200  0.447782319 0.530387721 0.564035422 0.650325632 0.675489154 0.737132908  145  Absorbance change vs. various conductivities of suspending media for M. smegmatis. Conductivity (μ S/cm)  Change in Absorbance (ΔA)  40 60 80 120 160 200 320 400 520 680  0.69849 0.64046 0.55219 0.80053 0.49356 0.47163 0.29066 0.23421 0.10762 0.05019  Various conductivities of suspending media vs. (k-ΔA)/ (2ΔA+k) for M.smegmatis. k  Conductivity (μ S/cm)  (k-ΔA)/(2ΔA+k)  2.23562  40 60 80 120 160 200 320 400 520 680  0.423148152 0.453618705 0.504022896 0.374043386 0.540551297 0.554910508 0.69044755 0.740152347 0.868263538 0.93554751  146  Appendix E. UBC Research Ethics Board's Certificates of Approval  147  148  149  150  151  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0058905/manifest

Comment

Related Items