DEVELOPMENT OF ELECTROSPUN NANOFIBER BIOSENSOR AND NUCLEAR MAGNETIC RESONANCE BAS ED BIOSENSOR FOR RAPID PATH OGEN DETECTION By Yilun Luo A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biosystems Engineering Š Doctor of Philosophy Electrical Engineering Š Dual Major 2020 ABSTRACT DEVELOPMENT OF ELECTROSPUN NANOFIBER BIOSENSOR AND NUCLEAR MAGNETIC RESONANCE BASED BIOSENSOR FOR RAPID PATH OGEN DETECTION By Yilun Luo Water and f ood contaminat ed with pathogen s cause millions of hospitalizations and thousands of deaths , and cost ly recalls in food and retail industries annually . Among them, the Shiga Toxin-producing Escherichia coli (STEC) causes foodborne outbreak s every year , which leads to more than 265,000 illnesses , 3,600 hospitalizations , and 30 deaths in the United States alone . Currently approved detection method s, such as culturing and colony counting , or polymerase chain reaction (PCR) , provide accurate diagnosis . However, these methods require long detection time (ranging from 6 to 24 hours ), high test ing cost, large -sized equipment , and/ or skilled personnel , limit ing their application in controlling outbreak s, reduc ing recall loss , or on -field diagnosis in developing countries . In this dissertation research , two biosensors were developed based on electrospun nanofiber and nuclear magnetic resonance ( NMR ) for rapid detection of STEC with high sensitivity . The electrospun biosensor was designed as lateral -flow immuno -sensor based on magnetic nanoparticles (MNPs ) and electrospun nanofibers . The MNPs were coated with conductive nano -shells and functionalized with antibody to extract target pathogen by immunomagnetic separat ion . Biocompatible nanofib rous membrane was synthesized by electrospinning technique , which was optimized for nano -porous structure and excellent capillary properties. The electrospun membrane was functionalized with antibody to capture the MNP -pathogen conjugate s by lateral -flow separation . As a result , the membrane™s conductivity was proportion al to pathogen concentration , which c ould be measured by a portable impedance analyzer. Owing to t he novel nanostructure , the surface area and mass transfer rate were significantly increase d. This improve d the biochemical binding effect and sensor signal to noise ratio. The biosensor™s sensitivit y limit was 61 colo ny forming units per milliliter ( CFU/mL ) and 104 cell culture infective dose per milliliter ( CCID/mL ) for bacterial and viral samples, respectively , with detection time of 8 min . The electrospun biosensor has advantages of low cost and high sensitivity, which can be used for on -field biodefense and food safety applications. In the second work, a portable nuclear magnetic resonance (pNMR) biosensor was developed based on antibody functionalized MNP s as proximity biomarke rs of the pathogen , which induce d micro -magnetic variation to accelerate NMR resonance signal decay . The pNMR was designed using a hand -held magnet of 0.47 Tesla , a high -power radio frequency (RF) transmitter , and an ultra -low noise receiver capable of detecting 0. NMR signal. The pNMR biosensor assay and sensing mechanism was used in detecting E. coli O157:H7 , and sensitivity limit was 76 CFU/mL in water samples and 92 CFU/mL in milk samples with detection time of 1 min. The pNMR biosensor is innovative for bacterial detection in food matrices and can be extended to other microbial or viral organisms by changing the antibod y specificity . Besides, the pNMR biosensor can be used for on-field healthcare diagnostic and biodefense applications owing to its advantages of portab ility and speed of detection . Copyright by YILUN LUO 2020 v This work is dedicated to my farther Xinhua Luo, my mother Zhongchun Xu, my wife Lei Jin, and my daughter Jessica Jinyi Luo. vi ACKNOWLEDGEMENTS I sincerely thank everyone who supported and contributed to the completion of this research throughout the years. First and foremost, I would like t o express my deepest gratitude to my advisor, Dr. Evangelyn Alocilja, for her continuous support and guidance over the years and for making everything possible. Her expertise, broad vision, and encouragement helped me finish my Ph.D. research. I would like to thank her for valuable advice, encouragement, and mentoring throughout my graduate study and research. I would also like to thank my co -advisor in the Department of Electrical and Computer Engineering, Dr. Timothy Hogan for his strong support and insig htful guidance during my dual -major doctoral research. I would like to thank the members of my Ph.D. guidance committee professors, Dr. Premjeet Chahal, Dr. Wei Liao, and Dr. Renfu Lu for their valuable advice, helpful discussions, and all the support. I thank all the members in the Biosensors research group, Michael Anderson, Barbara Cloutier, Yang Liu, Shannon McGraw, Sudeshna Pal, Emma Setterington, Edith Torres, Yun Wang, and Deng Zhang for sharing a creative, professional, and pleasant research envir onment. I also would like to express my appreciation to Dr. Lawrence Drzal and Steven Nartker in the Department of Chemical Engineering for the innovative research collaboration and valuable discussions. I especially would like to thank Dr. Nejadhashemi Am irpouyan, the Graduate Coordinator in the department, for his help and support to my doctoral readmission and study at Michigan State University. Besides, I would like to thank Hanna Miller, Michael Wiederoder, David Hochhalter, and all the other undergrad uate assistants who have worked with me over the past years. I also vii thank the Ann Arbor Technical Center and the Southwest Research Institute for all the support and assistance provided to complete this doctoral research work. Finally, I would like to exp ress my deepest gratitude and love to my parents, Xinhua Luo and Zhongchun Xu, and my family for their unconditional support, love, and believing in me during my Ph.D. pursuit. I truly appreciate my wife, Lei Jin, who accompanied me through the past years and provided countless encouragement and support no matter what difficulties or challenges she encountered and fought for. Ultimately, to our daughter, Jessica Jinyi Luo, who was born in the year before my dissertation defense. She brought us such precio us and memorable joy while transforming our lives and providing me with the best motivation for complet ing this dissertation. Yilun Luo viii TABLE OF CONTENTS LIST OF TABLES .......................................................................................................................... x LIST OF FIGURES ....................................................................................................................... xi KEY TO ABBREVIATIONS ....................................................................................................... xv CHAPTER 1 INTRODUCTION .................................................................................................... 1 1.1 Hypothesis ......................................................................................................................... 3 1.1.1 Contribution I: Electrospun Nanofiber Biosensor ............................................. 3 1.1.2 Contribution II: Portable NMR Biosensor ......................................................... 3 1.2 Research Objectives .......................................................................................................... 4 1.3 Research Significance and Novelty ................................................................................... 4 CHAPTER 2 LITERATURE REVIEW ......................................................................................... 7 2.1 Escherichia coli Pathogen ................................................................................................. 7 2.1.1 Traditional Methods of Detection .................................................................... 10 2.2 Biosensor based Detection .............................................................................................. 13 2.2.1 Acoustic Biosensor .......................................................................................... 14 2.2.2 Electrochemical Biosensor ............................................................................... 16 2.2.3 Optical Biosensor ............................................................................................. 17 2.2.4 Immunoassay Biosensor .................................................................................. 19 2.2.5 NMR based Microorganism Detection ............................................................ 21 CHAPTER 3 SYNTHESIS AND CHARACTERIZATION OF MAGNETIC NANOPARTICLES .................................................................................................. 26 3.1 Introduction ..................................................................................................................... 26 3.2 Materials and Methods .................................................................................................... 27 3.2.1 Sy nthesis of Conductive Immunomagnetic Nanoparticle ................................ 27 3.2.2 Synthesis of NMR Magnetic Nanoparticle ...................................................... 28 3.2.3 Magnetic Nanoparticle Characterization ......................................................... 29 3.3 Results and Discussion .................................................................................................... 30 3.3.1 Magnetic Nanoparticle Characterization and Synthesis .................................. 30 3.4 Conclusions ..................................................................................................................... 31 CHAPTER 4 BIOSENSOR BASED ON ELECTROSPUN NANOFIBERS AND MAGNETIC NANOPARTICLES FOR PATHOGEN DETECTION ........................................... 32 4.1 Introduction ..................................................................................................................... 32 4.2 Materials and Methods .................................................................................................... 34 4.2.1 Electrospun Material Synthesis ........................................................................ 34 4.2.2 Plasma Enhancement for Capillary Flow ........................................................ 38 4.2.3 Sensor Architecture and Detection Principle ................................................... 39 4.2.4 Test Pathogens and Antibodies ........................................................................ 44 4.2.5 Surface Functionalization ................................................................................ 45 ix 4.2.6 Immunomagnetic Separation ........................................................................... 47 4.2.7 Detection and Data Analysis ............................................................................ 49 4.3 Results and Discussion .................................................................................................... 51 4.3.1 Surface Functionalization with Antibody ........................................................ 51 4.3.2 Aligned Nanofibrous Membrane by Parallel Electrode Electrospinning ........ 54 4.3.3 Biosensor Detection ......................................................................................... 55 4.4 Conclusions ..................................................................................................................... 66 CHAPTER 5 BIOSENSOR BASED ON NMR AND MAGNETIC NANOPARTICLES FOR PATHOGEN DETECTION ...................................................................................... 69 5.1 Introduction ..................................................................................................................... 69 5.2 Materials and Methods .................................................................................................... 70 5.2.1 Design of Portable NMR Biosensor ................................................................ 70 5.2.2 Design of Portable NMR Duplexer .................................................................. 73 5.2.3 Design of Portable NMR Antenna ................................................................... 77 5.2.4 Design of Matching Networks of the pNMR Probe ........................................ 83 5.2.5 Evaluation of the Portable NMR Probe and the Matching Networks .............. 88 5.2.6 NMR Power Amplifier .................................................................................... 92 5.2.7 NMR Low Noise Amplifier ............................................................................. 96 5.2.8 NMR Detection using Magnetic Nanoparticle ................................................ 99 5.2.9 Effects of MNP on NMR Relaxation ............................................................. 103 5.2.10 Detection of pNMR Relaxation Time .......................................................... 115 5.2.11 Test Pathogen and Antibodies ..................................................................... 118 5.2.12 Magnetic Pathogen Separation .................................................................... 118 5.2.13 Sensor Architecture and Detection Principle ............................................... 119 5.2.14 Detection and Data Analysis ........................................................................ 120 5.3 Results and Discussion .................................................................................................. 121 5.3.1 Functionalization of Magnetic Nanoparticles with Antibody ........................ 121 5.3.2 MNP Antibody Functionalization .................................................................. 122 5.3.3 NMR R2 Relaxation Time .............................................................................. 123 5.4 Conclusions ................................................................................................................... 129 CHAPTER 6 CONCLU SION AND FUTURE WORK ............................................................. 131 6.1 Conclusions ................................................................................................................... 131 6.2 Recommen sdations for Future Work ............................................................................ 133 APPENDIX ................................................................................................................................. 135 BIBLIOGRAPHY ....................................................................................................................... 155 x LIST OF TABLES Table 2.1 Reported E. coli O157:H7 Outbreaks and Infections in the United States of the last five years (2016 to 2020) [16]. .............................................................................................. 8 Table 4.1 Dimension of the biosensor and its components .......................................................... 42 Table 4.2 P -value of biosensor test results (unaligned nanofiber) for E. coli O157:H7 ............... 61 Table 4.3 P -value of biosensor test results (unaligned nanofiber) for BVDV .............................. 62 Table 4.4 P -value of biosensor test results (aligned nanofiber) for E. coli O157:H7 ................... 63 Table 4.5 P -value of biosensor test results (nitrocellulose mat) for E. coli O157:H7 .................. 65 Table 5.1 Matching network design based on NMR probe characterization through resistance and inductance measurement .............................................................................................. 86 Table 5.2 RF performance of Henry Radio 50B HF power amplifier, including output power, frequency range, and operating mode. ......................................................................... 94 Table 5.3 RF performance of Miteq AU -1467 low noise amplifier, including frequency range, gain, and noise figure. .................................................................................................. 97 Table 5.4 RF performance of Sonoma Instrument 310 low noise amplifier, including frequency range, gain, and noise figure. ....................................................................................... 99 Table 5.5 P -value of NMR biosensor test results for E. coli O157:H7 in water samples ........... 127 Table 5.6 P -value of NMR biosensor test results for E. coli O157:H7 in milk samples ............ 128 xi LIST OF FIGURES Figure 2.1 Schematic diagram and working principle of a biosensor system. ............................. 13 Figure 2.2 The oscillating frequency is inversely proportional to the mass as indicated by the spring -mass oscillator system as an example. .............................................................. 14 Figure 2.3 Scheme of sensor operating principle for electrochemical biosensors (adapted and modified from [31]). ..................................................................................................... 17 Figure 2.4 Scheme of sensor structure and operating principle for optical biosensors (adapted and modified from [35]). ..................................................................................................... 18 Figure 3.1 Synthesis of conductive immunomagnetic nanoparticles by encapsulating gamma -iron -Fe2O3) nanoparticles with polymerizing aniline as conductive nano -shell. .. 28 Figure 4.1 Schematic diagram of electrospinning setup for nanofibrous membrane fabrication, consisting of the syringe, the needle, the nanofiber jet (the Taylor cone, the stable region, and th e instability region), and the rotating collector. ..................................... 35 Figure 4.2 Self -contained Nanofiber Electrospinning Unit (NEU, Kato Tech Co. J apan) using rotating drum as collector for high yield fiber fabrication. .......................................... 36 Figure 4.3 Scanning electron microscopy image of electrospun nitrocellulose nanofibers. ........ 37 Figure 4.4 Nanofibers spun across the gap of a parallel electrode collector on the NEU uni t ..... 38 Figure 4.5. The capillary flow capability comparison of electrospun nanofibrous membranes (A) without and (B) with the plasma en hancement. ........................................................... 39 Figure 4.6 Silver electrodes fabricated on electrospun nanofiber membrane using spray deposition method. ....................................................................................................... 40 Figure 4.7. Schematic of the biosensor structure and membrane assembly consisting of cellulose application and adsorption pads and electrospun cellulose nitrate capture pad. .......... 41 Figure 4.8. Detection scheme of the lateral flow immunosensor based on immunomagnetic nanoparticle and electrospun antibody functionalized capture membrane. ................. 43 Figure 4.9. The surface antibody functionali zation process for the electrospun biosensor capture pad. ............................................................................................................................... 47 Figure 4.10. The antibody functionalization process for the conductive M NPs. ......................... 49 xii Figure 4.11 Testing platform of electrospun MNP biosensor. (a) Test strip mounted on platform and sandwiched by copper elect rodes. (b) Design of the entire platform with three testing units .................................................................................................................. 50 Figure 4.12 The CLSM image of functionalized membrane with FITC antibody, (A) CLSM image of nitrocellulose nanofibrous membrane with FITC antibody functionalization, (B) CLSM image of nitrocellulose nanofibrous membrane without antibody. Significant fluorescence emission at 530 nm verifies the antibody immobilization. .. 52 Figure 4.13 FITC antibody functionalized electrospun membrane: (A) CLSM image verified that the fiber morphology retains after antibody functionalization. (B) Fluorescence image confirmed that antibodies are attached on membrane after wash step by significant fluorescent emission at 435 nm. ................................................................................... 53 Figure 4.14 The SEM and optical microscope image of functionalized membrane with FITC antibody (A) SEM image of the electrospun nanofibrous membrane, (B) optical microscope image of nanofibrous membrane and silver electrodes after antibody functionalization. .......................................................................................................... 53 Figure 4.15 SEM image of highly -aligned nanofibrous membrane synthesized by the para llel electrode collector electrospinning. ............................................................................. 54 Figure 4.16 The CLSM image at 600×: significant fluorescent emission verified strong ant ibody functionalization, and the highly -aligned fiber morphology still remained intact. ...... 55 Figure 4.17 SEM images of electrospun memb rane after test. E. coli O157:H7 were effectively captured on the functionalized fiber mat. ..................................................................... 56 Figure 4.18 SEM images of electrospun membrane after test. No bacteria were observed in the nanofiber mat without functionalization. ..................................................................... 57 Figure 4.19 Biosensor conductance signal versus test time of E. coli O157:H7 samples with different target concentrations: aligned nanofibrous biosensor ................................... 59 Figure 4.20 Biosensor (unaligned nanofibers) conductance signal versus test time of E. coli O157:H7 samples with different target co ncentrations: randomly oriented nanofibrous biosensor ...................................................................................................................... 60 Figure 4.21 Biosensor test results (unaligned nanofiber) for E. coli O157:H7 demonstrate a linear sensing response from 0 to 10 4 CFU/mL. .................................................................... 60 Figure 4.22 Biosensor test results (unaligned nanofibrous ma t) for BVDV virus demonstrate the linear sensor response from 10 1 to 10 3 virus dilution and control sample. .................. 61 Figure 4.23 Biosensor test results (aligned nanofibers) for E. coli O157:H7 demonstrated a linear relationship between resistance signal and bacterial concentration from 10 1 to 10 4 CFU/mL. The signal was significantly below the control from 10 5 to 10 7 CFU/mL. . 63 xiii Figure 4.24 Pathogen detection comparison of biosensors made of electrospun nanofibrous membrane and nitrocellulose mesoporous membrane. ................................................ 65 Figure 5.1 System Architecture of the Portable NMR, including FPGA control, modulation and demodulation system, and NMR sub -systems. ............................................................ 72 Figure 5.2 System prototype of the portable NMR, including a palm -sized permanent magnet, NMR transmitter, NMR recei ver, T/R switch, sample holder, and NMR probe. ........ 73 Figure 5.3 Complete set up of NMR coil antenna and duplexer switch for the NMR transmitter and receiver .................................................................................................................. 74 Figure 5.4 The crossed diodes pair shows high impedance for low voltages, and low impedance for high vot lages. The threshould is the built -in voltage. ............................................ 75 Figure 5.5 NMR coil antenna and duplexer switch based on quarter -wave impedance transformer: transmitter ON ......................................................................................... 76 Figure 5.6 NMR coil antenna and duplexer switch based on quarter wavelength transformer: transmitter OFF ............................................................................................................ 77 Figure 5.7 A high Q NMR antenna corresponds to higher selectivity, and is more immune to noise for the NMR receive r [153] ................................................................................ 82 Figure 5.8 The NMR probe, consisting of NMR coil antenna and a matching network, was designed to achieve high quality fact or to optimize NMR signal acquisition. ............ 85 Figure 5.9 NMR probe comprising of NMR coil antenna, NMR sample tube, tuning circuit, and non-magnetic probe holder is designed and fabricated in the Nano -Biosensors Lab. . 88 Figure 5.10 The NMR probe evaluation using a Key sight E5062 VNA for reflection coefficient and quality factor. Its matching network was tuned while measuring using the VNA to optimize the reflection coefficient ( S11) and achieved -23.6 dB. ................................. 89 Figure 5.11 The NMR probe was designed to achieve high quality factor to optimize NMR signal acquisition. Its matching networks was tuned to optimize the reflec tion coefficient using a VNA and achieved -23.6 dB at 19.918 MHz. .................................................. 90 Figure 5.12 Testing set up of the NMR transmitter using the func oscilloscope to calibrate the input power required for pNMR application. ................. 95 Figure 5.13 The input powe r to the NMR transmitter was calibrated to achieve 20 W output to the NMR probe. ............................................................................................................ 96 Figure 5.14 Testing set up of the NMR receiver u sing the function generator, attenuators, and the ....................... 98 xiv Figure 5.15 The Conventional nuclear magnetic resonance (NMR) spectrometer, Bruker 700 MHz NMR system [165]. .................................................................................... 100 Figure 5.16 The Zeeman effect: nuclei spin energies split when placed in an external magnetic field [168]. .................................................................................................................. 104 Figure 5.17 Dipolar interaction between nuclear spins, I and S, through space. The spin S induces a local field at the spin I. ................................................................................... 106 Figu re 5.18 NMR spin -echo technique to treat the inhomogeneous field of a small permanent magnet using 90 degree and 180 degree RF pulse trains. (adapted and modified from [176]) .......................................................................................................................... 116 Figure 5.19 NMR spin -echo technique and CPMG pulse sequence to detect the spin -spin relaxation time, T2 in a less homogeneous magnetic field. ........................................ 117 Figure 5.20 Working principle of the NMR based biosensor for pathogen detection ................ 119 Figure 5.21 Magnetic nanoparticles as biomarker to detect the target pathogen by the NMR measurement. ............................................................................................................. 120 Figure 5.22 TEM image and electron diffraction image (inset) of the MNPs. ........................... 122 Figure 5.23 NMR biosensor relaxation signal of detection: (A) control (blank) sample, and (B) bacterial sample .......................................................................................................... 124 Figure 5.24 The MNP -based pNMR biosensor™s measurement of relaxation time change, delta T2, for drinking water, which were contaminated by E. coli O157:H7. ..................... 127 Figure 5.25 The MNP -based pNMR biosensor™s measurement of relaxation time change, delta T2, for whole milk samples, which were contaminated by E. coli O157:H7 . ........... 128 xv KEY TO ABBREVIATIONS AC Alternative Current BVDV Bovine Viral Diarrhea Virus CEA Carcinoembryonic Antigen CDC Centers of Disease Control and Prevention CFU Colony -Forming Unit CLSM Confocal Laser Scanning Microscopy CMOS Complementary Metal Oxide Semiconductor CNT Carbon Nano Tube CPMG Carr -Purcell -Meiboom -Gill C4 Complement Component 4 DFB 1,4-difluorobenzene DI De-Ionized DMF Dimethylformamide ECL Electrochemiluminescent EGF Epidermal Growth Factor ELISA Enzyme -linked Immunosorbent Assay ENM Electrospun Nanofibro us Membrane FDA Food and Drug Administration FITC Fluorescein Isoth iocyanate FPGA Field Programmable Gate Array FSM Finite State Machin e xvi HUS Hemolytic -uremic Syndrome IC Integrated Circuit IgG Immunoglobulin G ITO Indium Tin Oxide LCP Lanthanide -complexed Polymer LFA Lateral Flow Assay LIBS Laser-Induced Breakdown Spectroscopy LOC Lab -On-Chip MNP Magnetic Nanoparticle MPN Most Probable Number MWCNT Multi Walled Carbon Nano Tube NEO Neomycin NIAID National Institute of Allergy and Infectious Disease NV Nitrogen Vacancy NMR Nuclear Magnetic R esonance PANI Polyaniline PBS Phosphate -buffered Saline PCR Polymerase Chain Reaction pNMR Portable Nuclear Magnetic Resonance PVDC Polyvinylidene Chloride QCM Quartz Crystal Microbalance QNS Quinolones Antibiotics RF Radio Frequency xvii SEM Scanning Electron Microscope SERS Surface -enhanced Raman Scattering SNR Signal to Noise Ratio SPR Surface P lasmon Resonance STEC Shiga Toxin-producing Escherichia Coli SWCNT Single Walled Carbon Nano Tube TEM Transmission Electron Microscopy THF Tetrahydrofuran TMP Trimethyl Phosphate VHDL Very High Speed Integrated Circuit Hardware Description Language WHO World Health Organization 1 CHAPTER 1 INTRODUCTIO N Clean drinking water is essential to public health and safety . Pathogenic contamination s cause enormous medical expenses , massive loss es to retailor s and food industry , and serious thr eats to bio security . Escherichia coli O157: H7 , an important waterborne pathogen, is associated with acute and lethal illness, including diarrhea, hemorrhagic colitis, and hemolytic -uremic syndrome (HUS), which has attracted extensive attention from academia and health agencies . E. coli O157: H7 infects human alimentary tract, produces Shiga toxin, and induces abdo minal cramps with hemorrhagic diarrhea [1] . Studies from 10 of the 14 subregions of the World Health Organization (WHO) indicates that the global breakout of E. coli is 2.8 million cases annually . In the United State alone, estimates of E. coli O157:H7 induced hemorrhagic colitis cases are 63,000 per year [2] [3] . E. coli grow rapid ly at exponential rate. Hence, r apid and accurate testing of water samples are crucial to identify the pathogenic bacteria and control the spread of disease . The current standard detection method s are growth based , which requir e 18 hours or longer for the identification [4] [5] . Culture plating -based methods detect the bacteria through metabolic pathways, not by pathogenic capabilities. These methods are slow and have limitation in selectivity among related species [6] . Testing method s based on DNA gene sequences are applied to the detection and ident ification of bacteria , including polymerase chain reaction (PCR), isothermal amplification, and bio -barcode method. PCR directly detects the gene sequences and is highly sensitive in identification . However, it require s special ized equipment, strict reagent storage , and highly s killed personnel , and hence often limited to laboratory application [7] . Isothermal amplification is less expensive than PCR owing to simplified equipment requirement s but is still 2 limited due to the requi rement f or enzy me storage at low temperatu re [8] . The bio -barcode method is based on DNA amplification and is applicable for enzyme free detection . However, it involves DNA input and the detection process is time consuming [9] . Antibody detection is based on its recognition of antigenic proteins on the cell surface of the target bacteria. It is used in bio -barcode detection and enzyme -linked immunosorbent assay (ELISA) [10] . ELISA has higher sensitivit y, but it is expensive and not suitable for high throughput measurement. Its detection tim e often tak es 24 hours or longer . This dissertation describes the research work on two biosensors developed for the detect ion of E. coli O157: H7. The first part of the dissertation is the development of lateral flow immuno -biosensor based on electrospun nanofibers and magnetic nanoparticles ( MNP s). The second part of the dissertation describes the development of a n MNP -based portable nuclear mag netic resonance ( NMR ) biosensor . The dissertation is organized as follows. Chapter 2 presents a review of the literature on technologies relevant to this research work , including the current standard testing, biosensor -based detection and NMR based detection . Chapter 3 describes the generation of MNPs synthesized by amine -functi onalized iron oxide using thermal decomposition method , and t heir optimization for the electrospun biosensor and the pNMR biosensor application s. Chapter 4 describes the development of electrospun nanofibrous membrane , surface modification, antibody functionalization , and the design for MNP -based lateral flow immun o-assay biosensor . Chapter 5 summarizes our efforts to develop NMR sensing of MNP -labeled pathogen by combination of an improved cell binding process, portable NMR development and performance optimization, and 3 NMR detection based on selective immuno -assay MNP . Chapter 6 contains the summary of the research and recommendations for future work. The following section describes the research hypothesis, ob jectives, and novelty of the dissertation work performed. 1.1 Hypothesis The two research contributions presented in this dissertation are based on the following hypotheses: 1.1.1 Contribution I: Electrospun Nanofi ber Biosensor An antibody -based biosensor platform can be developed for rapid detection of pathogenic contaminants through impedance readout of an MNP -pathogen sandwich assay captured by lateral flow separation using the electrospun membran e. 1.1.2 Contribution II: Portable NMR Biosensor A portable NMR biosensor platform can be developed to measure the concentration of target pathogen in water and food samples based on the acceleration of proton spin relaxation induced by antibody conjugated MNP s. 4 1.2 Research Objectives The overall objective of this dissertation research is to develop rapid pathogen detection biosensor s to detect E. coli O157:H7 . The detailed objectives of this project are as follows. To synthesize MNPs , functionalize them with antibody, and optimize them for impedance detection and NMR detection , respectively To synthesize electrospun nanofibrous membran es and optimize them for lateral flow immuno assay to capture target pathogen s and separate supernatant To develop a lateral flow immuno assay biosensor based on MNP and electrospun nanofibrous membrane for rapid detection through electrical impedance measurement To develop a portable NMR biosensor for rapid testing of water and milk samples by measuring perturbation s of proton spin relaxation caused by MNPs immuno -conjugated on the target pathogen 1.3 Research Significance and Novelty The novelty of the research contribution presented in this dissertation lies in exploring the use of an innovative nanomaterial , nitrocellulose electrospun nanofibro us membrane (ENM) , in function s of lateral flow separation, pathogen capture , and biosens ing transducer. High throughput fabrication method , capillary enhancement, immuno functionalization, and impedance sensing are innovat ive . The adaptability of nano biosensor has been demonstrated in detecting both bacteria l and viral pathogen s. Current literature shows a sizable amount of research on nitrocellulose ENM synthesis. However, the capillary action enhancement and lateral flow assay (LFA) application have not been reported in prior literature . New literature shows ENM s based on metal oxide or 5 DNA network have been implemented as impedimetric biosensor s to detect chemical compound s, ion s, and cancer cells [11] -[14] . However, as of today , this is the first report of a n ENM in an integrated design of LFA and impedimetric immuno assay as one biosensor system , and successful detection of both bacteria l and viral pathogens . The electrospun lateral flow biosensor demonstrated excellent performance in sensing response and detection speed. Owing to high surface -to-volume ratio and unique nanostructure , the ENM enhanced capillary action , assay kinetics, and immunochromatography ability compared to conventional membrane material. This led to improved target pathogen binding effects , filtration, and separation from sample supernatant. The electrospun biosensor is of low cost and small size, and capable of rapid detection with high sensitivity , which can be used for on -field application of food safety, water monitoring , and biodefense. The innovati on of the research contribution in the second biosensor includes exploring the use of polyaniline maghemite MNP , filtration assay to quantify the pathogen by MNP concentration, as well as inexpensive and high signal -to-noise ratio ( SNR ) NMR probe and RF transceiver for microbial detection in complex matrices . The biosensor is novel in aspects of design and application. The versatility of the pNMR biosensor was demonstra ted in bacterial detection in both water and dairy food samples. Current literature shows a number of p ortable NMR system s for noninvasive imaging . Recent literature shows sever al biosensing applications of micro NMR s that are fabricated based on integrated circuit (IC) and micro -fabrication technolog ies . Their detection of biotarget s, such as cancer cells and molecules , are based on immuno -clustering of MNPs. The reported sensitivity is in the range of 103 CFU/mL (Staphylococcus aureus ). The fabrication 6 approach is also more costly. To the best of our knowledge, this is the first report of NMR biosensor with successful detection of bacterial pathogen in water and dairy food samples . The sensitivity demonstrated a lower limit and wider linear response, from 101 to 10 4 CFU/mL . To minimize infectious disease spread and reduce costly food recall s and medical expenses , it is crucial to implement rapid pathogen detection system s with high sensitivity and on -field capability . The micro NMR systems in the existing literature are small in size and suitable for on -field application. Based on the water and food monitoring application, t here is still a need to improve their sensitivity limit s and sample volume ( currently up to sub - ). We developed a portable NMR biosensor based on high SNR probe and NMR transceiver , resulting in a higher sensitivity and wider linear sensing response , as shown in Figure 5.24 and Figure 5.25. It was also capable of testing sample s with a volume up to 180 L, allowing combined detection of multiple sample s in one test , which is essential to rapidly handle massi ve test sample s during a pandemic. With the novel filtration assay, the test sample could be easily concentrated to handle larger test volume or to detect lower concentration. The pNMR detection is nondestructive, which allows further investigation through other testing methods. The system is low in cost and highly portable , which is favorable for on-field and rapid detection of foodborne pathogen s. 7 CHAPTER 2 LITERATURE REVIEW 2.1 Escherichia coli Pathogen With incidences occurr ing every year , microbial contaminations in water and food safety have become main concer ns. One of the leading bacterial pathogens is Escherichia coli O157:H7 , which belongs to the Shigatoxigenic group of E. coli (STEC) . It is a gram -negative rod -shaped bacterium of micro size around 1.1 to 1.5 µm in diameter , 2.0 to 6.0 µm in length , and with a cell volume of 0.6 to 0.7 µm3. The i nfection can be caused by ingestion of contaminated food or water, or oral contact with contaminated surfaces . The resultant illness can last for 5 to 10 days with symptoms including severe and acute hemorrhagic diarrhea, abdominal cramps, headache, vomiting, and nausea . Infection with some patients, p articularly children five years and younger , immu nity -compromised person and the elderly , can lead to life -threatening hemolytic uremic syndrome (HUS) , which cause red blood cells to breakdown , or be destroyed, and kidney fail ure [15] . The occurrence rate is around 2% to 7% of infections. In the United States, the main cause of acute kidney failure in children is the HUS, most of which are caused by E. coli O157:H7 infectio n. Meanwhile, E. coli O157:H7 is one of the major bacterial pathogens accountable for foodborne illness and product recall , such as drinking water, beverages , fresh vegetables, and fast food, etc. There were at least eight confirmed food -linked outbreaks of the bacteria in the U.S. in the last five years ( 2016 to 2020 ) as reported by the United States Centers of Disease Control and Prevention ( CDC) (Table 2.1) [16] . In 2018, food contamination of E. coli O157:H7 in r omaine lettuce caused two foodborne infectio n outbreak s. The first one lasted from March to June and 8 caused a total of 210 cases across 36 states with 96 hospitalization and 5 death. The second one lasted from October 2018 to January 2019, and caused a total of 62 cases across 16 states with 25 hospitalization and a recall of all 18 kind s of sandwiches and other food products distributed in 5 states , leading to a health alert issued by the CDC [17] . In the latest outbreak as recent ly as January 2020, E. coli O157:H7 contaminated fresh salad kits caused infections in multiple states and costly hospitalization. Furthermore , the bacteria can contaminate various other kinds of food products , such as spinach, sprouts, flour, cookie dough, butter, cheese, hazelnut , sausage, beef and pork , etc., which have been reported by the CDC. Table 2.1 Reported E. coli O157:H7 Outbreaks and Infections in the United States of the last five years (2016 to 2020) [16] . Outbreak Time Source of the Outbreak Reported Cases States Hospitalizations Recall 2019/11/05 to 2020/01/15 Fresh Express Sunflower Crisp Chopped Salad Kits 10 5 4 No 2019/9/20 to 2019/12/21 Romaine Lettuce 167 27 85 Yes 2018/10/07 to 2019/01/09 Romaine Lettuce 62 16 25 Yes 2018/03/13 to 2018/06/28 Romaine Lettuce 210 (5 death) 36 96 No 2017/11/05 to 2018/01/25 Leafy Greens 25 (1 death) 15 9 No 2017/01/04 to 2017/05/04 I.M. Healthy Brand Soy Nut Butter 32 12 12 Yes 2016/06/27 to 2016/10/19 Beef, Veal, and Bison products 11 5 7 Yes 2016/01/17 to 2016/03/25 Alfalfa Sprouts 11 2 2 Yes 9 As estimated by the CDC, the totally annual cost of E. coli O157:H7 infections is around $271,418,690 [18] . The yearly total cases of illness are estimated at 63,153 . 11,737 cases require d the assistance of physicians , of which 1,806 patients were hospitalized . Among those, 10 cases were deadly even though HUS was not developed , while 302 cases did developed HUS with 12 acute death s and 10 premature death s at a later point. The infection s result in various costly expenses to patient s, employers, and the health care system for medical expen se, acute and premature death loss, wage loss , and productivity loss . Furthermore , the E. coli O157 is highly effective in multiplication and infection. It can spread in water, food, soil, or on surfaces that has been contaminated by body fluids or feces of animal or human. The bacteria can live in the intestines of healthy cattle and other animals without illness. Study by the United States Food and Drug Administration (F DA) indicates that fewer than 10 to 100 colony forming units (CFU) is sufficient to cause infection [19] . It is required by the World Health Organization (WHO) that E. coli bacteria must not be detectable in any 100 -ml water sample to verify the quality and safety of water intended for drinking or public distribution [20] . Furthermore, E. coli O157:H7 is classified as a fiCategory Bfl, the second highest priority, pathogen for biodefense by the CDC and the National Institute of Allergy and Infectious Disease (NIAID), due to its ease of wide spread in water, beverages, and va rious food sources, and highly virulent illness consequences [21] [22] . As a result , it is not tolerant for any level of E. coli O157:H7 in water or food intended for human use . And it is essential to identify micro quantities of this pathogen in unknown sample s effectively and rapidly . The standard method of E. coli O157:H7 detection require s highly trained personnel 10 and complex instrument with long testing tim e. The identification involves several steps including enrichment in selective media, incubation on differential agar to isolate s orbitol non -fermenting colonies , biochemical identification , and definitive identification , lasting around four days . Detection by PCR is capable of phenotypical and serological identification with high accura cy. However, it require s long testing time lasting several days, and high cost due to complex sample preparation steps, such as DNA extraction and amplification. For faster testing , the real -time PCR method can be used to detect the bacteria , after the selective enrichment process , to provide a testing result of positive or negative within 24 h and are not applicable for on -field applications [23] . However, it still require s the culture method and PCR to confirm presumptive positive results for three days [19] . Hence , it is still of high demand to improve detection time and sensitivity for E. coli O157:H7 for disease surveillance, prevention strategies, and water and food monitoring . The following sections described the traditional methods as well as novel methods based on biosensors for pathogen detection used in water and food monitoring efforts. 2.1.1 Traditional Methods of Detection The conventional methods of bacteria l identification are through gas generation or colorimetric change , of which the determination is based on the metabolic differences. Among this category, there are three techniques approved by the EPA, including presence or absence testing, multiple fermentation tube, and mem brane filtration. The presence/absence testing technique is the first option for bacterial testing [24] . The procedure involves adding a known volume of water sample, typically 100 mL, to a liquid or powder media 11 in a sterile container, and determination by change of color after incubation of 24 hours. This technique has simple procedures of testing and ana lysis and does not require incubator owing to its wide temperate range. However, its results are not quantitative, and is not appropriate for testing known contaminated samples. The multiple fermentation tube technique is performed by growing the test sam ple in a liquid medium with a second tube inversely placed inside to show gas generation [25] . The culture medium suitable for coliforms can be lactose, lauryl trypto se, or lactose bile broth. The test sample are incubated inside the tube under anaerobic conditions at 35-45°C for 24 -48 hours to allow gas generation . The presence of the coliforms is indicated by pH or turbidity or change of the culture medium, or by the gas presence. The result is reported as a most probable number (MPN) index, which is determined by comparing the pattern of tube numbers showing growth at each dilution (the positive results) with statistical tables. Hence, it is a statistical estimate of the starting concentration with unit of MPN index per 100 mL, but not an actual count of bacteria presented in the sample. A second fermentation for 24 hours is required for positive quantification of fecal coliform to determine total and fecal coliform q uantities. This method remains being the key conventional procedure for sample matrices that are non -transparent or colored such as milk or turbid field samples with semi -solids. The membrane filtration technique integrates the filtration concentration a nd growth assay. It starts with filtering a measured volume of liquid sample under vacuum through a cellulose acetate membrane of uniform pore diameter, typically 0.45 µm. The bacteria captured on the membrane is incubated in a sterile container using a se lective medium, such as m -Endo -type agar to form red 12 bacteria l colonies after growth at 35 °C of 24 hours. Other medium that can be used for this technique includes Lactose agar with Tergitol 7 , Teepol broth , ChromAgar, MacKonkey agar , MacKonkey -sorbital agar , and Rainbow agar . This technique is regularly used in monitoring water systems owing to its capability of testing relatively large numbers of samples by filter concertation. It also requires less testing time than the multiple ferm entation tube technique and can give a direct count of total and fecal coliforms present in the sample. However, the result can still be affected due to uncountable plate growth, over -crowding, and capture of non -target microbes that can out -compete the coliforms . Culture media and buffered dilution water may be prepared in the field, but this requires the transport of all necessary equipment, which may include measuring cylinders, beakers, distilled water, autoclavable bottles, a large pressure -cooker and a gas burner or other source of heat. It is possible to prepare the c ulture media on field, but needs transport all necessary equipment, including distilled water, autoclavable bottles, beakers, measuring cylinders, a large pressure -cooker and a heat sourc e, such as gas burner . The EPA approved methods have been used for many years in water sample testing and monitoring. However, a major drawback is their long testing time with common methods taking 24 to 48 hours, and rapid methods taking 12 hours. There is still a considerable gap with the actual requirement indicated in the previous section to meet the demand of reducing disease infection and costly recall. Besides, the growth system can be affected by contamination with other microbials, which is usuall y common in on -field samples. Non -target microbials can over -compete and slow the growth of the target, leading to false negative results. Moreover, selective culture medium cannot guarantee the optimal growth conditions thus giving false negative results when the target pathogen is stressed or of low concentration. 13 2.2 Biosensor based Detection Biosensors are sensing device s that combine a bioreceptor element and a physicochemical transducer to detect biochemical substances. The system structure and working principle of a biosensor is demonstrated in Figure 2.1. The bioreceptor element is a biologically derived or biomim etic component that captures and recognizes the sensing target through highly specific binding or other biochemical integrations. Once recognized, the physicochemical signal generated by the biochemical recognition reaction is converted by the transducer into o ther forms, which can be amplified and processed for detection. The bioreceptor is versatile in its sensing elements, which can be composed of antibodies, aptamers, cell receptor, enzymes, molecularly imprinted polymers, microorganisms , nucleic acids, and whole cells, etc. According to different sensing principles, biosensors can be categorized into many forms, including electrochemical, optical, mechanical, and magnetic biosensors, which are discussed in detail in the following part of this section. Figure 2.1 Schematic diagram and working principle of a biosensor system. Antibodies Nucleic Acids Aptamers Enzymes Whole Cells Electrochemical Optical Acoustic Magnetic NMR Amplifier Data Acquisition Analyte Bioreceptor Transducer Signal Processing 14 2.2.1 Acoustic Biosensor The acoustic biosensor is based on piezoelectric crystal, a unique material capable of transforming mechanical vibration into electrical energy, and vice versa . During the sensing, an alternative current (AC) is applied to excite standing wave in the crystal. The resultant resonant vibration has a characteristic fr equency, which is highly sensitive to the surface properties of the crystal. Hence, if the crystal is functionalized with a bioreceptor, once the sensing target is captured, it will cause a shift in the resonant frequency due to the mass change as shown in Figure 2.2. Figure 2.2 The oscillating frequency is inversely proportional to the mass as indicated by the spring -mass oscillator system as an example. The mass change has a linear relation with the frequency shift as indicated by the Sauerbrey equation ( Equation 1) [26] , which can be used to quantify the target concentrat ion. m0 m1 k k f0 f1 The o scillation frequency, f, decreases as the mass, m, increases 15 =2.3×10 Equation 1 where, is the frequency change in Hz, is the resonant frequency of the crystal in MHz, and A is the coating area in cm 2, and is the captured mass in grams. Acoustic immunosensors have been reported, of which immuno assay is functionalized on the crystal surface as the sensor™s bioreceptor. Muratsugu et al. developed a q uartz crystal microbalance (QCM) biosensor using antibody specific to human serum albumin to form a label free assay. It successfully detected albumin in urea (albuminuria) in a range 0.1 Œ100 µg/mL [27] . In another research , a QCM immunosensor was developed by Deng and coworkers to detect a protein in human , the complement component 4 (C4) [28] . The working electrodes were modified with nafion membrane and then by the target -specific antibody. Detection range of 0.08 Œ1.6 µg/mL and relative standard deviation around five percent were achieved [28] . Funari and coworkers reported another QCM immunosensor, of which spatially oriented antibodies against gluten were functionalized on gold electrodes. I t reached sensitivity limit of 4 ppm for the gluten and detection range between 7.5 and 15 ppm [29] . In summary, the acoustic biosensor can be of low cost and durabl e since piezoelectric crystals are abundant, inexpensive, and robust, which is suitable for biosensor applications under challenging physiochemical conditions with low cost. In addition, the acoustic biosensor provides great flexibility, wide dynamic sensi ng range, and is capable of label free detection. The acoustic biosensor has a limit ation in detecting low molecular weight analytes because they cause lower decrease of oscillation frequency. Besides, the sensitivity is not ideal for detecting large analy te, 16 such as microbial cell , because it does not act as an ideal mass point and only the portion in the proximity of the bio receptor is involved in frequency change of oscillations [30] . 2.2.2 Electrochemical Biosensor The electrochemical biosensors convert chemical information from bioreceptor events into various form of electrical signals, such as potentiometry, amperometry, imp edance, capacitive, or conductometry. The working principle and system structure of the electrochemical biosensor is demonstrated in Figure 2.3 [31] . It measures the conductance of ions or electrons in a sample solution or fibrous network using inert electrodes, direct or alternating current, and an alternating null current. The electrode is a major component , which is responsible for the binding of target biomolecules and the transport of electron /ions . Label -free application has been developed based on different electrochemical sensing techniques, such as impedance spectroscopy for DNA detection using single walled carbon nanotubes (SWCNTs) as support for DNA probe [32] , and electrical conductometry using graphene support with r GO/AgNP composites [33] , etc . For labelled application, binding molecules, such as antibodies, enzymes, or aptamers, can be used to conjugate signal probes on the sen sing target to further amplify or transduce the binding event in to signals to be easily measured and quantified . Recently, n ew material has been studied including carbon -based nanomaterial, such as SWCNT, multi -walled CNT (MWCNT), and graphene; and non -carbon nanomaterials, such as metallic nanoparticles, porous silica, nanowire, indium tin oxide (ITO), and organic polymers, etc. A ratiometric electrochemical biosensor was developed by Cai et al. using Polythionine ŒGold (PTh ŒAu) as an electrode. It can dete ct a tumor marker , 17 carcinoembryonic antigen (CEA) , with good specificity within a wide linear range, and a detection limit of 2.2 pg/mL [34] . Owing to the large surf ace area to volume ratio and unique mechanical and electron transport properties, synerg ic improvements are achieved in sensing performance and analytical sensitivity by enhanced loading capacity and increased mass transport of reactants. Figure 2.3 Scheme of sensor operating principle for electrochemical biosensors (adapted and modified from [31] ). 2.2.3 Optical Biosensor The optical biosensor utilizes an optical transducer to detect the target analyte under study. Various types of bioreceptors adopted in other biosensor platforms can also b e used in their optical counterparts, including antibodies, antigens, enzymes, nucleic acids, receptors, whole cells and tissues. The optical transducer then detects the recognition events by measuring the evanescent field in close proximity to the biosens or surface in terms of surface plasmon resonance (SPR), evanescent wave fluorescence , or optical waveguide interferometry . The working principle and system structure of the electrochemical biosensor is demonstrated in Figure 2.4. 18 Figure 2.4 Scheme of sensor structure and operating principle for optical biosensors (adapted and modified from [35] ). The sensing of the optical biosensor relies on the number of biorecognition and optical transducing events. Hence, changing material structure or increasing s urface area and porosity to enhance the amount of radiation reflected or emitted are beneficial to improve sensitivity. Nano materials, such as nano particles and nano fibers, are promising to achieve such improvements owning to their versatility and uniqu e nanostructure. Optical biosensors are relatively new, and their importance is expected to grow in healthcare and pathogen detections. They can enable large -scale high -throughput testing of multiple samples simultaneously. However, optical biosensors fo r general practical application are still under development. Their current applications are still limited to academic and pharmaceutical environments. 19 2.2.4 Immunoassay Biosensor Immunoassay biosensor adopts antibodies as the bioreceptor with advantage of the hi gh selectivity provided by the molecular recognition and binding between antibody and antigen. The immunoassay biosensor can be generally categorized into two types: label -free and labeled. The label -free application relies on detecting the physical change s directly induced by the antibody -antigen complex. The labeled application adds a sensing label to the antibody -conjugated target, forming a sandwich structure to amplifier or transduce the original conjugation event. Their transducers have a variety of sensing forms, such as optical sensing including luminescent, fluorescent, and surface plasmon resonance, electrochemical sensing including voltammetry, amperometry, and impedance spectroscopy. Owing to the continuous advancement of instrument electronic s, the size and cost of the sensor transducer have been reducing over the years and is technically ready for portable applications. The bioreceptor remains the crucial part to be further improved on sensitivity and detection time to fulfil the requirements for water and food testing . Enhancing the surface area of the biosensor is an effective approach through maximizing the number of reaction sites. Hence, nanomaterials of electron nanofibrous membrane and magnetic nanoparticle are the promising solution th at provides these advantages, owing to their excellent surface to area ratio and unique nanostructures. Immunoassay biosensor has recently been developed based on nanofibers, polymers, nanoparticl es, CNTs, and graphene [36] -[40] . Fellows and coworkers developed a rapid lateral flow assay (LFA) based on streptavidin -conjugated gold nanoparticles (AuNP s) as reporter molecules to screen single -stranded DNA aptamers for the detect ion of a glycoprotein , cluster of differentiation 4 (CD4). The sensor was able to detect 250 ng of human CD4 in 9 min [41] . A 20 lateral flow immunoassay (LFI) to detect proteins was developed by Qiu and coworkers utilizing antibody functionalized CNTs as a colored (black) tag [42] . The capture antibodies on the test zone of LFI captured target proteins , and then immobilized the CNTs labelled by detection -antibodies, resulting in a black colored line on the LFI to enable visual detection of protein . It could detect rabbit IgG in spiked human plasma in 20 min with sensitivity limit of 1.3 pg /mL [42] . Zheng and coworkers developed a label -free immunosensor based on electrochemiluminescent (ECL) to detect -Trophin protein [43] . AuNPs were linked on the in dium -tin oxide subtracted via (3-aminopropyl) trimethoxysilane based polymer to enhance ECL and immobilize the -Trophin antibodies. The detection time was approximately 2 hours. The detection limit of -Trophin was 1.26 ng/mL[43] . Shi and coworkers developed a portable lateral flow assay (LFA) biosensor to simultaneously detect neomycin (NEO) and quinolones antibiotics (QNS) based on immuno -nanoprobe s and surface -enhan ced Raman scattering (SERS) detection [44] . The LFA achieved sensitivity limits of 10 ng/mL for NEO and 200 ng /mL for NOR using visual detection , and 0.37 pg/mL and 0.55 pg /mL using SERS [44] . Gondhalekar and coworkers developed a laser -induced breakdown spectroscopy (LIBS) for lateral -flow immunoassays (LFIAs) to detect E. coli using labels of AuNP or lanthanide -complexed polymers (LCPs) [45] . The LIBS system was applied on a commercial LFIA (based on nitrocellulose membrane s) to detect AuNP labeled E. coli . It achieved a sensitivity limit of 8.89 × 103 CFU/mL in approximately 3 hours of total testing time [45] . Compared to chemical synthesis methods of nanomaterial s, the electrospinning technology has unique advantages to fine tune fiber properties with ease, such as fiber compositions, orientation, diameter, and length, etc. The capability of the fibrous membrane, including surface area and capillary action, can be optimized for binding, filtration, and signal transducing to improve biosensor performance. 21 2.2.5 NMR based Microorganism Detection Nuclear magnetic resonance (NMR) is a physical phenomenon of which atomic nuclei absorb and reemit elec tromagnetic radiation at certain frequency when placed in a magnetic field. NMR technique exploits this phenomenon by analyz ing the magnetic properties of atomic nuclei to determine physical and chemical properties of biomolecules. Since its sensing signal is able to pass through turbid raw samples, simplif ying sample preparation processes and sav ing analysis time , NMR techniques have wide applications in non-contact and non -destructive biomedical and food diagnoses [46] . In standard testing , NMR analys is were performed in laboratories using a strong stationary magnet , in which the sample of interest was placed inside [47] . Although the system produced accurate results, the magnet and instruments were rather expensive , heavy , large, and unportable [47] [48] . This makes the NMR technique less useful and less practical for on -site applications. However , continuous development has led to the advancement of portable NMR (pNMR) detection hardware and sens ing performance. pNMR sensors were developed using magnetic nanoparticles and microparticles as proximity sensors to amplify molecular interactions [49] . The ir sensing is based on the reversible self -assembly of dispersed magnetic particles into stable nano -assemblies [49] . When a few magnetic nanoparticles conjugate to their molecular target through affinity ligands, they form magnetic clusters which result in a corresponding decrease in the bulk spin -spin relaxation time ( T2) of its surrounding water molecules [50] . Their measurements can be performed on turbid samples with simplified sample preparation . Furthermore, the sensing is faster than those by surf ace -based techniques , which reli es on molecular diffusion of targets to the sensing elements. These advantages make the proximity assay ideal for fast, simple and high -throughput sensing applications , especially in miniaturized device format. Such type of sensors is smaller in size and suited for on -site and field applications . They 22 are also cheaper to manufacture and less costly to maintain as compared to the bulky conventional NMR systems . A portable NMR system for a noninvasive spin -echo imaging of livin g plants in their natural environment was reported in 2000 [51] . Recent advances in micro -fabrication technologies have accelerated the development of palm -sized NMR and high -throughput NMR transceivers . The palm -sized NMR was based on integrated circuit (IC) technique and was tested in detecting biological molecules and cancer cells by measuring the sample™s NMR relaxation time which was inversely proportional to the immuno -clustering of magnetic nanoparticles [50] . The reported sensitivity for detection of cancer cell s was in the range of 10 3 to 104 cells/mL [52] . Other portable NMR developments include different designs [47] -[49] [51] [53] [54] . Hash and coworkers developed a n NMR biosensor to detect Vibrio paraha emolyticus bacteri um spiked in shrimp tissue s [55] . The detection started with DNA extract ion from test sample s. Iron nanoparticle s coated with target -specific biomarkers were applied to bind with V. parahaemolyticus DNA, which could be detected by molecular mirroring NMR technology [56] . The NMR biosensor output spin -spin relaxation time, T2, which correlat ed with the quantity of the V. parahaemolyticus DNA. The biosensor c ould detect different quantities of V. parahaemolyticus DNA, which was equivalent to cell concentrations ranging from 105 to 10 8 CFU/mL in approximately 1 hour of testing time [55] . An on-chip probe -based portable NMR was developed by Gupta and coworkers to detect a malaria parasite, parasitaemia Plasmodium falciparum in human blood [57] . The system utilized a permanent magnet of 0.5 Tesla to detect the NMR signal of water proton . The signal frequency was 21.287 MHz. The NMR probe was designed combining a pl anar circular coil antenna and a matching network (a L-section circuit with a shunt capacitor and a series capacitor ). After capacitor tun ing, the quality factor of the probe was optimized and measured as 24.407, using the same measurement method as described in Section 5.2.5 . The NMR T2 23 difference s between test sample s of control (blank) and parasitaemia were measured to determine the target concentration. The NMR biosensor achieved a sensitivity limit of 0.0001% parasitaemia , the percentage of infected red blood cells used to monitor the infection progress and the patient recovery. Th is was at the same level as their PCR result s and was lower than the other methods that they conducted for comparison, including blood smear, fluorescence activated cell sorting , SYBR Green staining [57] . Gossuin and coworke rs studied using low -field (0.5 Tesla and 1 Tesla) NMR T2 sensing to detect the synthetic malaria pigment -hematin based on its own paramagnetic effect on water proton (without adding superparamagnetic MNPs for signal amplification ) [58] . It could detect -hematin with concentration of 3.88 mg/ml. The result indicated that paramagnetic particles were more difficult to detect by NMR than the superparamagnetic particles . The paramagnetic effect -hematin on relaxation decay was not enough to accurate ly detect malaria without the use of a large -fiel d magnet, T2 (not T1), constant temperature, or any preliminary sample preparation, such as microcentrifugation [58] . Lu and coworkers developed a low -field microfluidic NMR device (0.443 Tesla) to detect tumor markers using immunomagnetic nanoparticles (IMNPs) [59] . A multi -layer microfluidic NMR probe (probe diameter of 1.7 mm ) was designed for sample transport and detection . The matching network was of the same design as described in Section 5.2.5. The device utilized a commercially available electronic control system (Bruker minispec console) as the NMR tran sceiver. The transverse relaxation time change T2 was measured to determine th e target concentration in approximately 1 hour . It detected three biomarkers, respectively , and achiev ed sensitivity limit of 10-1 ng/mL and linear response from 100 ng/mL to 5×102 ng/mL for i mmunoglobulin G (IgG ), sensitivity limit of 5×10 -1 ng/mL and linear response from 100 ng/mL to 102 ng/mL for m ucin glycoprotein 1 ( MUC1 ), and sensitivity limit of 101 cells/mL and linear response from 5×10 2 cells/mL to 5×10 4 cells/mL for human breast 24 adenocarcinoma cell line (MCF -7) [59] . Janis and coworkers developed a novel two-dimensional NMR sensor with a microfluidic diamond quantum sensor [60] . The microfluidic chip transported the sample through a medium low field (1.5 Tesla Halbach magnet ) for prepolari zation , and subsequently teste d it inside using a low field ( B0 = 13 mT, Helmholtz coils) . The NMR sensor utilized optically probed nitrogen -vacancy (NV) quantum defects in diamond , exited by a linearly polarized 532 -nm green laser beam , to detect NMR signals with high -spectral resolution from micron -scale sample volumes . The NV NMR signal was detected using a custom -built epifluorescence microscope . It achieved 2D correlation spectroscopy of liquid analytes : trimethyl phosphate (TMP), or 1,4-difluorobenzene (DFB ), within an effective detection volume of app roximately 40 picoliter in a spectral resolution of 0.65 ± 0.05 Hz [60] . The development of a high -throughput NMR spectrometer used complementary metal oxide semiconductor (CMOS) technology to integrate an array of high sensitivity micro -coils with interfacing rad io-frequency circuits on the same chip [61] . A micro nuclear was also developed to miniaturize the sensing system to be palm -siz e with a portable sub -Tesla magnet, and electronically automate for multi -step and multi -sample chemical/biological diagnosis [67] . The utilized microfluidic and microelectronic technologies to enable the coordination between the droplet management detection . Targets in sub - water samp les, captured by specific probe -decorated magnetic nanoparticles, were sequentially quantified by their spin Œspin relaxation time (T2 relaxometer that can fit in a 2 mm by 2 mm silicon chip was also reported [54] . In order to minimize the pathogen infection and costly product recall, rapid, sensitive, and portable detection of E. coli O157:H7 is crucial in applications of healthcare and food supply. Studying 25 from the basic NMR design by Fukushima and Roeder [151] , we developed a portable NMR (pNMR) and biosensor assa y that can rapidly and sensitively detect foodborne pathogens. Our pNMR biosensor is novel in filtration assay, use of high signa l-to-noise ratio but inexpensive NMR probe, and RF transceiver for microbial detection in complex matrices. The proximity bioma rker uses an antibody -functionalized magnetic nanoparticle (Ab -MNP). The system is low cost and does not require skilled operators. It has high testing throu ghput, small size and high porta bility, which is suitable for in-field foodborne pathogen detection . 26 CHAPTER 3 SYNTHESIS AND CHARACTERIZATION OF MAGNETIC NANOPARTICLES 3.1 Introduction Magnetic nanop article s (MNPs) have promising properties in a wide variety of applications , which continue to attract research focus in recent years [61] -[65] . MNP s exhibit no net magnetization due to the random thermal flipping of magnetic moments but can be effectively manipulated by external magnetic field s. In addition, i t can be functionalized by different affinity ligands, such as aptamers, lectins, folic acid, and epidermal growth factor (EGF) , etc . [1] [16] . MNPs can be conjugated to target cells or directly absorbed inside the cell for magnetically controlled non-destructive cell manipulation [69] and for lab -on-chip (LOC) applications [70] . MNPs have wide applications in medical diagnostics and treatments. After being manipulated or delivered to target tissue/cell, n anoparticles can be heated by an external alternati ng magnetic field as an experimental cancer treatment called magnetic hyperthermia [71] . MNPs can be used to target specific cells or tissues inside human body in or der to accurately deliver drugs with optimal quant ity [72] . MNPs can also be used to magnetically manipulate free -floating cancer cells to be carried out of the body for lab analysis [73] . For sensor applications, MNPs can be used to extract target pathogen or biomaterial from supernatant solution as the magnetic separation for sample purification and concentration [74] [75] . Electr ochem ical detection using magnetic nanoparticles has been developed to form immunoassa y with target pathogen, and measure cyclic voltammetry (CV) signal s or electrical conductance signal s [76] [77] . 27 Magnetic nanoparticles synthesis has been established by different chemical process including : co-precipitation method [78] , microemulsion method [79] and flame spray synthesis method [80] to achieve different core and shell material /composition , particle size, magnetization, and electrochemical properties . This chap ter describes synthesis methods of two magnetic nanoparticles, which were developed in the lab , for conductance biosensor and NMR biosensor application s, respectively . 3.2 Materials and Methods 3.2.1 Synthesis of Conductive Immunom agnetic Nanoparticle The conductive immunomagnetic nanoparticles were synthesized from -Fe2O3) nanoparticles as core, and polymerized aniline as shell, which was biologically and electrically activated by acid dopi ng [1] [82] [83] . The -Fe2O3 (maghemite ) nanoparticles , which were purchased from Sigma -Aldrich (St. Louis, MO) , were disper sed in a mixture of 50 ml 1M HCl, 0.4 ml of aniline monomer and 10 ml de -ionized water , and sonicated at 0°C for l hour to disintegrate the cluster . Then a slow drop -wise addition of the oxidant, ammonium persulfate ((NH 4)2S2O8), at a rate of 0.1 ml/min was added to this solution mixture with constant stirring . The solution color change d from rust brown to dark green which indicated the formation of the conductive emeraldine form of polyaniline (green) coating on the smaller -Fe2O3 nanoparticle s (brown) . This reaction continued for 4 hours with constant stirring at 0°C. Finally, the green solution was filtered and washed with 1M HCl, and 10% methanol and diethyl ether. The filtered product was dr ied for 18 hours at room temperature . The synthesis of conductive immuno magnetic nanoparticles is demonstrated in Figur e 3.1 below. The polymerization and acid doping process 28 were determined to obtain a uniform nanoparticle size and optimized magnetisms and conductivity to be used for the electrospun latera l flow biosensor. Figure 3.1 Synthesis of conductive immuno magnetic nanoparticles by encapsulating gamma -iron oxide (-Fe2O3) nanoparticles with poly merizing aniline as conductive nano -shell . 3.2.2 Synthesis of NMR Magnetic Nanoparticle The NMR biosensor detects target pathogen concentration by measuring the sample ™s 1H NMR relaxivity , which is determined by magnetic field alternation from the pathogen and magnetic nanoparticle conjugation . The small er and more uniform nanoparticle with higher magnetism is capable to increase conjugation ratio per pathogen and enhance the whole conjugation™s magnetization , which is critical to the sensitivity of NMR biosensor. -Fe 2O3 + Aniline HCl (NH 4)2S2O8 Oxidative Polymerization Conductive Magnetic Nanoparticle Polyaniline Iron Core +H3N 29 Hence , for NMR application , amine -functionalized Fe3O4 (magnetite) magnetic nanoparticles (AMNPs) were sy nthesized separately in the lab based on thermal decomposition of Fe -chloride method [85] , with improvement modifi cations . First , 1.08g of iron ( III) chloride hexahydrate (FeCl 3.6H2O), 2g of sodium acetate and 7 ml of ethylenediamine were added in 30 ml of ethylene glycol for 2 h at room temperature to obtain a homogenous golden yellow mixture . The solution was transferred into a teflon -lined stainless -steel pressure vessel (Parr Instrument Company , Moline, I L) to be sealed and heated at 200° C for 15 h for thermal decomposition dehydration and Fe3O4 reaction . After cool ing to room temperature, the synthesized nanoparticles were thoroughly cleaned by magnetic separation of three times of 20 ml water wash and three times of 20 ml ethanol wash. During each clean step, samples were placed inside a strong permanent magnet to remove the supernatant. Finally, the synthesized magnetic nanoparticles were dried under vacuum for 12 h. 3.2.3 Magnetic Nanoparticle C haracterization The synthesized conductive immuno magnetic nanoparticle and NMR amine nanoparticle have been characterized to evaluate particle size, conductivity and magnetic capability [82] . A high performance TEM, JEOL 2200 FS , with field emission cathode and acceleration voltage of 200 kV was used to investigate particle structur e, size , and content uniformity . Both particles™ m agnetic properties were measured using a superconducting quantum interference device , MPMS SQUID (Quantum Design Inc., CA ). And h ysteresis magnetization was measured by cycling magnetic field from +15 kOe to Œ15 kOe, at constant temperature of 300 K. A four -point probe , Pro -4 (Lucas/Signaton Corp., CA) , was used to evaluate polyaniline (PANI ) particle™s electrical 30 conductivity at room temperature , which uses separate pairs of current -carrying and voltage -sensing electrodes to increase accuracy . 3.3 Results and Discussion 3.3.1 Magnetic N anoparticle Characterization and S ynthesis The electrical conductance of the conductive PANI magnetic nanoparticle was measured by four -point probe measurement s to be 3.3 S/cm (+/- 0.0 4 S/cm , n = 3) at room temperature , whereas -Fe2O3 nanoparticles indicated 3.4 × 10-5 S/cm (0.17 S/cm standard deviation , n = 3) conductivity under the same measurement . The high conductivity of the synthesized PANI magnetic nanoparticles confirmed the insulator -conductor transition into the electrically active polyaniline. At 300 K temperature , the conductive PANI magnetic nanoparticles exhibited saturation magnetization of 44.1 emu/g measured at a magnetic field of 15 kOe using a SQUID magnetometer , while un -Fe2O3 nanoparticles showed 64.4 emu/g measured under the same conditions . The magnetization reduction can be explained by a bipolaron conduction mechanism that caused conductive polyaniline to be diamagnetic [84] . However, the nanoparticles™ magnetization is more than sufficient for the intended biosensor application of immuno magnetic separation . Transmission electron microscopy (TEM) studies confirmed the effective coating of PANI around the -Fe2O3 particles. T he synthesized conductive PANI MNPs indicated a controlled diameter ranging from 50 to 100 nm, whereas the -Fe2O3 nanoparticles from Sigma -Aldrich (St. Louis, MO) had an average diameter of 20 nm (4.6 nm standard deviation ). It qualifies as NP because it meets the crite ria of diameter between 1 and 100 nm [86] . Different weight ratio s of -Fe2O3 to aniline of 1/0, 1/0.1, 1/0.2, 1/0.4, 1/0.6 , and 1/0.8 were evaluated and finally determined to be 1:0. 6 to optimize both size and conductivity . The 31 lower weight ratio resulted in MNPs with lower electric conductance ( 0.768 S/cm) , while the higher ratio led to mostly amorphous shape and smaller saturation mag netization ( 33.5 emu/g ). The conductive PANI MNPs and the amine NMR MNPs both exhibited excellent superparamagnetic behavior , which can be uniformly distributed in solution or became highly magnetized when exposed to external magnetic field. As reported in the lab, the amine -functi onalized NMR particles were highly mono -dispersing with diameter of 25 nm, which was verified by TEM imagin g. The saturation magnetization significantly increased to 80 emu/g at 300 K, compared to PANI MNPs. Due to advantages in both particle size and magnetization, more amine MNPs can be effectively conjugated with the target pathogen each with stronger magnetism , which is important to NMR biosensor sensitivity. 3.4 Conclusions In this chapter, c onductive PANI magnetic nanoparticles and amine -functio nalized NMR magnetic nanoparticles were successfully generated with sizes ranging of 50 to 100 nm and 20 to 30 nm respectively. Particle size and conductance were optimized for electrospun lateral flow biosensor application. Size and magnetization were opt imized for NMR biosensor . The particles were successfully functionalized with antibodies and implemented for the biosensors for pathogen detection in C hapter 4 and 5. 32 CHAPTER 4 BIOSENSOR BASED ON ELECTROSPUN NANOFIBERS AND MAGNETIC NANOPARTICLES FOR PATHOGEN DETECTION 4.1 Introduction Biocompatible material s, such as nitrocellulose, polyvinylidene fluoride , and polyether sulfone , are proven for their excellent binding capability and can capture biomolecul e and pathogen cells to be separate d from bio -samples. Low cost and high sensitivity biosensor s have been developed based on immunoassay or immunochromatography for infectious pathogen detection [87] , ovulation monitoring , drugs and chemical analy sis in different applications [88] , including veterinar y testing [89] , agricultural and environmental monitoring , and product quality evaluation [90] . Nanomaterials are promising to fabricate biosensors with high sensitivity , rapid response , and low cost , owning to their unique structure and biochemical properties. In recent years, resear ch progress has been made in develop ing new nano biosensors to enhance surface area and reduc e cross section, thereby providing a more effective immobilization to capture bio -target s [91] [92] . Most of the development work is focused on using nanomaterial s with two -dimensional or three -dime nsional structure d mesoporous layers. New sensor material s with one -dimensional (1D) structure s have been introduced , such as carbon nanotube and metal oxide nanowires , providing unparalleled capability of rapid mass transfer for analyte molecules [93] . The 1D nanostructure has a significant advantage in surface to volume ratio , which leads to an excellent binding effect and sensor response. However, sensors compose d of single nanowire s or nanofiber s suffer from 33 sens ing variation due to their difference in size and diameter , which is difficult to control in nanofabrication. On the other hand, sensor s consist ing of nanowire /nanofiber networks can effectively average the response of each individual nanomaterial, providing advantages of reliable performance, low noise, and high reproducibility and repeatability . Electrospinning is a versatile , high -throughput, and cost -effective technology to produce the nanofiber networks, or nanofibrous membrane with control led and uniform fiber diameter in nanoscale [94] . Its fabrication relies solely on the use of high -voltage electrostatics to extract ultrafine solid threads from the material solution without the need for coagulation chemistry , high temperatures , or high pressure . Owning to novel nanostructure , the surface area of the resultant nanofiber is significantly increased compared to that of the conventional planar material, thereby enhanc ing both biochemical reaction rate and target binding effect [95] . In addition, the electrospinning technology provides many attractive advantages for the development of high -performance nanomaterials for biosensor applications, such as inherent stability, high yield, low cost, and compatibility with other micro fabrication process es [96] . For biomedical applications, such as tissue engineering [97] , drug delivery [98], and artificial organ implant s [99] , electrospinning methods have been developed to fabricate na nofibers with biomimetic structures in different morphologies . Moreover, electrospun nanomaterial s also demonstrated improved performance in molecular absorption [100] and cell adhesion [101] , owning to their enhanced binding capability. In past years, their application s for biosensors ha ve attracted extensive attention s and demonstrated promising testing results , such as biomolecular detection and enzyme functionalization [102] -[104] . However, electrospun biosensor s for whole cell detection of microbial or viral pathogens ha ve not been reported in prior literature. 34 This chapter presents a new electrospinning method to synthesize nitrocellulose nanofibrous membrane , and its optimization and functionaliz ation techniques of biolog ical treatment for biosensor application. A new biosensor based on electrospun capture membrane s has been developed . It integrates multiple techniques including magnetic separation, capillary immunoassay, and direct charge electrical measurement to achieve rapid and quantitative detection of whole cell bacteria and vir us. The electrospun nanofibrous membrane was synthesized from nitrocellulose polymer solution . The process condition s and parameters of the electrospinning were optimized to obtain aligned fiber networks. The membrane was treated with a plasma to further improve capillary action and a surface antibody functionaliz ation method was developed for the membrane and optimized for the target pathogen . The result ing membrane has a n ultrafine fiber diameter of about 150 nm . The high surface area of the material enable d more bio -reception events to occur , facilitate d lateral flow assay kinetics , thereby enhancing pathogen capture and separation from supernatant. Finally , the electrospun biosensor was tested using bacterial and viral inoculated samples to evaluate the binding and separation performance of the novel bio -modified nanofibrous membrane. 4.2 Materials and Methods 4.2.1 Electrospun Material Synthesis The electrospinning process uses electrostatics repulsion effect generated by high voltage power source between the needle spinneret and collector to produce an ultra -fine nanometer polymer jet delivered by a syringe pump from a polymer composite material liqui d, as shown in Figure 4.1. 35 Figure 4.1 Schematic diagram of electrospinning setup for nanofibrous membrane fabrication , consisting of the syringe, the needle, the nanofiber jet (the Taylor cone, the stable region, and the instability region ), and the rotating collector . The electrospun nanofibrous membrane was fabricated using the Nanofiber Electrospinning Unit (NEU, Kato Tech Co. Japan) , which is illustrated in Figure 4.2 below . The NEU device™s high voltage power source can be adjusted from 0 to 39 kV, and self -contained for safety. The generated electrospun nanofiber was collected on its metal roller collector, which is capable to produce a uniform fiber membrane . Syringe High Voltage DC Power Source Collector Polymer Solution Needle Rotation Taylor Cone Nanofibers Instability Region Stable Region 36 Figure 4.2 Self -contained Nanofiber Electrospinning Unit (NEU, Kato Tech Co. Japan) using rotating drum as collector for high yield fiber fabrication. The electrospun biosensor membrane was fabricated using n itrocellulose polymer for its excellent biocompatibility and solubility in common solvent . The electrospinning process was controlled by the self -contained NEU device. Its fabrication conditions were optimized as follows , in order to control nanofiber diameter optimal for bio -detection . The nitrocellulose polymer of 8 wt% was dissolved in a mixed solvent system consisting of 60% tetrahydrofuran (THF) and 40% dimethylformamide (DMF) , which provide s optimal viscosity and surface tension for electrospinning [109] . A 20 ml syringe pump with 18-gauge needle was used to extrude the polymer solution at 0.2 ml/h. The applied high voltage between the needle and collector was tuned to be 10 kV and the distance was determined to be 6 cm . The polymer solution was electrospun on polyvinylidene chloride (PVDC) substrate into unwoven nanofibrous membrane . The fiber membrane quality wa s verified using Scanning Electron M icroscope (SEM ) imaging , which is demonstrated in Figure 4.3 below. High Voltage Collector 37 Figure 4.3 Scanning electron microscopy image of electrosp un nitrocellulose nanofibers. This electrosp un fabrication system is capable to synthesize reproducible nano fibro us membrane made of 3D layers of smooth and defect -free nitrocellulose polymer nanofibers with uniform diameter of approximately 150 nm. When usin g the rotating drum collector, its synthesized electrospun fibers were typically in a random orientation, which limiting their capillary performance. A single parallel electrode pair used as collector enabled production of highly aligned mats to further en hance the capillary capability as shown Figure 4.4 below [111] . This unwoven membrane was uniform in fiber deposition depth and maintained surface and bulk chemi cal compositions which were expected for this material . Electrospun Nanofiber 38 Figure 4.4 Nano fibers sp un across the gap of a parallel el ectrode collector on the NEU unit 4.2.2 Plasma Enhancement for Capillary Flow The IMNP electrospun biosensor utilized capillary flow and magnetic nanoparticle antibody conjugation to separate and detect biological target . In order to enhance capillary capability , surface nitrate groups of the electrospun nanofiber mat were removed by O 2 plasma using 120 W RF plasma at 13.6 MHz with O2 at 250 mTorr , which was verified using X-ray Photoelectron Spectroscopy ( XPS ) spectrum an alysis [109] . After plasm a enhancement, the material property changed from hydrophobic to hydrophilic, and was verified by capillary c ontact angle test. As shown in Figure 4.5 below, before plasma treatment, contact angle between water droplet edge and nanofiber mat was 135°, which indicates that the surface was highly hydrophobic with low surface energy (>12 0°). After the plasma treatment, it changed to be 56° and indicated that the 39 electrospun mat became hydrophilic , which is good for capillary flow biosensor application (< 90°). Figure 4.5. The capillary flow capability comparison of electrospun nanofibrous membrane s (A) without and (B) with the plasma enhancement . 4.2.3 Sensor Architecture and Detection Principle A conductometric lateral flow biosensor based on the electrospun material was designed to consist of three porous membranes: sample application pad, capture pad, and absorption pad [82] . The electrospun nitrocellulose membrane was implemented as the capture pad due to its excellent biocompatibility, surface area to volume rat io and capillary properties. The fiber surface was bio -modified for antibody attachment usin g glutaraldehyde (CH 2(CH 2CHO) 2) as cross linker. A metal mask with pattern of parallel rectangular electrode pair was covered on top of the electrospun membrane. The colloidal silver ink was airbrush sprayed to fabricate a uniform silver electrode pair wit h gap distance of 0.5 mm . The mask protects the area between the electrodes from the corrosive silver ink solvent, which becomes the lateral flow channel of the capture pad as shown 135 ° 56 ° H2O H2O Nanofiber mat Nanofiber mat 40 in Figure 4.6 . The sprayed electrode has uniform pattern and consistent resistance determined to be 1 . The chemical composition was carefully chosen to prevent the erosion of nanofibers. Figure 4.6 Silver electrodes fabricated on electrospun nanofiber membrane using spray deposition method. The cellulose membranes with flow rate of 180 ml/min (Millipore, MA, USA) were used for the application and absorption pads due to their excellent filtration and sopping properties [110] . The application pad was used to control the flow of the sample onto the capture pad. The excess solution was absorbed in the absorption pad which modulated the capillary action. After cleaning with sterilized and deionized (DI) water, the application and absorption pads were cut into 7 × 5 mm2 and 10 × 5 mm 2, respectively. Capillary experiments were performed to optimize the Silver Ele ctrode Electrospun Nanofiber 41 membrane size for flow rate and sample volume. The larger pad size s resulted in slow er lateral flow across the application and absorption pad s (> 20 min ) and even partial flow stopped at the absorption or capture pad. The smaller pad size s led to sample overflow on one or all three pad s and unbound MNP to remain on the capture pad , which was verified in control (blank) sample testing through microscop y imaging . The optimized capture membrane had dimension of 23 × 5 mm2 for a test sample of 100 µL. The overall dimension of the IMP electrospun biosensor was 40 × 5 mm 2 (Table 4.1). Finally, the biosensor was assembled by attaching the three membrane pads onto a polyvinylidene c hloride ( PVDC ) substrate via polystyrene adhesive backing as shown in Figure 4.7 below. The biosensor unit was connected to a data acquisition system li nked to a computer via copper wiring substrate to measure the resistance of the biosensor which would indicate the target pathogen concentration (National Instrument, TX, USA). Figure 4.7. Schematic of the biosensor structure and membrane assembly consisting of cellulose application and adsorption pads and electrospun cellulose nitrate capture pad. 40 mm 5 mm 0.5 mm Adsorption Pad Capture Pad Silver Electrode Application Pad 42 Table 4.1 Dimension of the biosensor and its components Biosensor zone Dimension (mm) Application 7 × 5 Capture 23 × 5 Absorption 10 × 5 Overall 40 × 5 The pathogen detection principle of the biosensor is demonstrated in Figure 4.8. The antibody -modified conductive MNPs were mixed and incubated with the test sample to conjugate with the target pathogen . Then, the test sample was purified using magnetic separation proces s for 3 times [82] by holding the magnetically susceptible conjugation complex close to an external magnet to extract the supernatant by pipette. The purified sa mple , which was also conductive ly label ed by the MNP s, was dispensed on the application pa d and initiated the lateral flow . The porous structure of the application pad modulated the flow and filter ed large size d impurities. During the flow, the target pathogen was conjugated by the antibody in the electr ospun membrane and captured in the capture pad . The non -target biomolecule and ex cess magnetic nanoparticles in the solution were subsequently removed by the capillary action and absorbed in the adsorption pad. Finally, when the entire process reached equilibrium, the conductive MNPs captured in the electrospun membrane were proportional to the target concentration since the sample volume is constant . Thus, the presence and concentration of pathogen could be determined by measuring the conductan ce signal of the membrane via its two silver electrodes , as shown in Figure 4.19 in the following section 4.3. 43 Figure 4.8. Detection scheme of the lateral flow immunosensor based on immunomagnetic nanoparticle and electrospun antibody functionalized capture membrane. 44 4.2.4 Test Pathogens and Antibodies 4.2.4.1 Escherichia coli O157:H7 A pure culture of Escherichia coli O157:H7 was obtained from the collection of the Nano -Biosensors Laboratory (Department of Biosystems and Agricultural Engineering, Michigan State University). To make a stock culture, E. coli O157:H7 test strain s were inoculated using sterile loop into 10 mL of Tryptic soy nutrient broth from the Difco Laboratories (Detroit, MI) and incubated for 24 h at 37 °C . The 24 h stock culture was serially diluted in 0.1% peptone water in logarithmic scale to obtain different concentrations from 101 to 107 colony fo rming units per milliliter (CFU/mL). Each dilution was used as a test sample during the succeeding biosensor detection experiments. The antibodies used for E. coli O157:H7 biosensor were an affinity purified goat anti -E. coli O157:H7 polyclonal antibodie s from KPL, Inc . (Gaithersburg, MD), and purified mouse anti -E. coli O157:H7 monoclonal antibodies from Meridian Life Science, Inc. (Saco, ME). 4.2.4.2 Bovine Viral Diarrhea Virus In addition to the bacterial pathogen detection, t he biosensor was also evaluated to detect a viral pathogen, Bovine Viral Diarrhea Virus (BVDV) . It is of genus Pestivirus of the family Flaviviridae , including BVDV1 and BVDV2 . The virus consists of an envelope and a nucleocapsid in spherical shape . It has na no size with diameter ranges from 40 nm to 60 nm [113] . The BVDV not only infects bovine , but also various breeds of domestic and wild ruminants and pigs . In acute infections , BVDV can result in respiratory and reproductive symptoms, and enteric 45 symptoms , such as diarrhea and almost always immune suppression, causing the animal vulnerable to secondary infections. Some strains of BVDV leads to persistent infections without active symptoms for a long time . It is of greater harm as a major source to infect BVDV to the herds. Hence, t he United States Department of Agriculture (USDA) urges the need of effective surveillance programs to reduce BVDV infections in the national heard, and reliable methods to detect both persistent and acute infections. The USDA states that firobust field -ready tests that both detect and differentiate viral pathogensfl is needed in order to achieve this [114] . Samples of bovine viral diarrhea virus (BVDV) in serum were collected from BVDV -infected cattle maintained by the Department of Large Animal Clinical Sciences at the Michigan State University . These BVDV samples were stored at -80 °C before use. A BVDV test solution was thawed and serially diluted four times to get 10 -1, 10 -2, 10-3, and 10 -4 dilutions prior to use. Each dilution was used as a test sample during the succeeding biosensor detection experiments. The antibodies used for the BVDV biosensor were affinity purified mouse anti -BVDV 15c5 (gp48) monoclonal antibodies (Ed DuBovi, Cornell University, Ithaca, NY) and purified swine anti -BVDV polyclonal antibodies (USDA National Animal Disease Center , Ames, I A). All the experiments were performed in a certified Biological Safety Level II laboratory. 4.2.5 Surface Functionalization For pathogen detection application, the electrospun membrane was functionalized with the polyclonal antibody to act as capture pad of th e biosensor , as illustrated in Figure 4.9. To further enhance protein binding and reduce pH effect, glutaraldehyde (C 5H8O2) was used as a cross -46 linking agent to attach antibodies to the electrospun nano fiber membrane . The concentratio n and quantity were determined 2 of nanofibrous membrane . Less glutaraldehyde (10 and 15 ) led to inadequate antibody binding to the member , which was verified by the confocal laser scanning microscopy imaging described in Section 4.3.1 . First, glutaraldehyde solution was evenly dispensed on the membrane , and then allowed to dry in a biosafety cabinet for 10 minutes. After soaking with a phosphate buffer solution (PBS ) buffer to remove excess solution , 0.5 mg/ml polyclonal antibodies (affinity purified goat anti -E. coli O157:H7 polyclonal antibod ies or purified swine anti -BVDV polyc lonal antibod ies ) were dispensed on the electrospun membrane with a concentration of 10 µg/in 2, and incubated at 25 °C for 30 minutes. The process was completed after a wash ing step using tris buffer with 0.1% tween -20 to remove excess chemical and block unbound antibody. The tween -20, or polysorbate 20, is a polysorbate -type nonionic surfactant formed by the ethoxylation of sorbitan , which is used to saturate binding sites on surfaces . The immuno -functionalized nanofiber membrane was stored at 4°C in a refrigerator until needed (up to about 1 week) . 47 Figure 4.9. The surface antibody functionalization process for the electrospun biosensor capture pad. The antibody attachment protocol was verified using fluorescein isothiocyanate (FITC) labeled antibody and confocal laser scanning m icroscopy (CLSM) to measure the amount of em ission of fluorescent antibody . The significant emission at 530 nm optical wavelength confirmed an efficient antibody attachment on the electrospun membrane . Besides, CLSM images also demonstrated that the novel nanostructure by the electrospinning method was retained after immuno -surface functionalization. 4.2.6 Immunomagnetic Separation The conductive magnetic nanoparticles were synthesized from aniline monomer and gamma iron -Fe2O3) nanoparticles usin g a sol -gel chemical solution deposition method [82] . The Polyclonal antibodies to E. coli Glutaraldehyde Crosslinker O=CHCH 2CH2CH2HC=O + + Y ŠNH2 Y ŠNH2 Antibody attachment to capture pad Y ŠNH=CH(CH 2)3HC=HN Š Y Washing & Blocking Functionalized Capture Pad O=CHCH 2CH2CH2HC=O Electrospun nanofiber mat 48 polymerized MNPs™ polyaniline (PAN I) shell was transformed to be electrically conductive by acid doping. For E. coli antibody functionalization, 2.5 mg of the conductive PAN I magnetic nanoparticles were dispensed for uniform dissolution of affinity purified mouse anti -E. coli O157:H7 monoclonal antibody (2.5 mg/mL) was added to the soluti on for conjugation , according to the tuning process developed in our lab [115] . For BVDV antibody functionalization, the conductive PAN I magnetic nanoparticles dispensed in 1 mL PBS were sonicated for 10 minutes , and -BVDV 15c5 (gp48)) in PBS was added . The mixture solution was then incubated at 25 °C in a hybridization oven rotating at 30 rpm for 1 h our. As a result, t he antibody, which is negatively charged, was conjugated with the positively -charged PAN I shell of the nanoparti cle s by electrostatic force [81] . After incubation, the antibody functionalized magnetic nanoparticles (AFMNs) were cleaned by magnetic separation to remove impurities . The magnetic separation wash was performed three times with a blocking buffer consisting of 100 mM Tris ŒHCl buffer (pH 7.6) and 0.1% (w/v) casein (Figure 4.10). After the final wash step, the antibody functionalized magnetic nanoparticles were re-suspended in 2.5 mL 0.1M PBS a nd stored in a 4 °C refrigerator until needed. 49 Figure 4.10. The antibody functionalization process for the conductive MNPs . For immunomagnetic separation , the AFMNs were vortex mixed and placed in a magnetic separator for 2 minutes. The supernatant was discarded using a micropipette. Then 1 mL of pathogen sample solution obtained by serial dilution was added and incubated at 25 °C in a hybridization oven rotating at 30 rp m for 30 minutes. After three magnetic separation rinses (by adding 1 mL of tris buffer solution with 0.1% casein and removing the supernatant), the AFMNs which captured the target pathogen were finally re-suspended in 1 mL of PBS. As a result, the test sample solutions with different concentrations were purified by magnetic separation and labeled with electrical conductance, and ready for biosensor detection . 4.2.7 Detection and Data Analysis The biosensor testing unit consisted of a plastic printed circuit b oard (PCB) base and imprinted parallel copper electrodes pair, which was connected to the input of a data acquisition system as shown in Figure 4.11 below . The paral lel copper electrodes were 25 mm apart to mount the electrospun MNP biosensor in between. A pair of secondary copper electrodes, which appear as Monoclonal Antibody to E. coli O157:H7 Antibody Functionalized EAPMs + Y Y Y Y Y Y Y Y EAPMs Incubation Magnetic Separation 50 flaps on the base electrodes , in Figure 4.11 (a), was placed over the silver electrodes of the biosensor™s capture pad to form a low resistance sandwich structure . This ensure d an excellent electrical contact to the capture pad to directly measure conductance of the conjugated MNP s in nanofiber membrane . The biosensor testing unit was designed to include multiple measurement units , in Figure 4.11 (b), to simultaneously measure three electrospun MNP biosensors. Figure 4.11 Testing platform of electrospun MNP biosensor. (a) Test strip mounted on platform and sandwiched by copper electrodes. (b) Design of the entire platform with three testing units For signal measurement, the biosensor strips were connected via the testing unit to a data acquisition (DAQ) syste m, NI USB -6221 (National Instrument, Austin, TX) . The DAQ system was linked to a computer with USB interface and controlled by graphical user interface software in LabVIEW (National Instrument, Austin, TX) . For each sample, after immunomagnetic separation, a volume of 100 µL (E. coli O157:H7 ) of the test sample solution (200 µL for BVDV) was applied to the application pad of the biosensor and the conductance signal across the silver a). b). 51 electrodes was measured by the DAQ system. When capillary flow equilibrium was achieved at around 8 min, the biosensor resistance across the silver electrodes was recorded by the DAQ system to determine target concentration . The sa mple lateral flow time from the application membrane to the capture membrane was approximately 1 min . Conductance data was recorded 30 sec before test sample application and continued through 8 min until capillary flow equilibrium can be achieved . For data analysis, a minimum of three replications were performed for each experiment. All biosensors were calibrated using a control sample which consisted of the same AFMN s suspended in sterile DI water except pathogen . Standard deviations and mean values for the data of each experime nt were calculated. Statistical analysis was performed based on a single factor analysis of variance using SAS ANOVA. 4.3 Results and Discussion 4.3.1 Surface Functionalization with Antibody The effect of immuno surface functionalization was verified using the confocal laser scanning microscopy (CLSM) method . The f luorescein isothiocyanate (FITC) labeled antibody (530 nm and 435 nm ) was functionalized on th e electrospun nanofibrous membrane using the same immuno -functionalization process described in Chapter 4.2.5 above. After two additional wash ing process using tris buffer with 0.1% tween -20 to rem ove unbound antibody , CLSM was used to image the functionalized membrane , and measure its laser excited fluorescence wavelength to confirm antibody attachment . As shown in Figure 4.12 A and Figure 4.13 A, signif icant fluorescent emission at 530 nm (green) and 435 nm (red) were observed , respectively , which confirmed a strong antibody attachment to the electrospun membrane. The untreated nanofibro us membrane was also analyzed to compensate background noise generated by the substrate and 52 fibers ( Figure 4.12 B and Figure 4.13 B). In addition, the CLS M image (Figure 4.12 A and Figure 4.13 A) and SEM image ( Figure 4.14 A) also demonstrated that the unique nanostructure of electrospun membrane was retained after the bio -modification process, which is critical for biosensing performance. Figure 4.12 The CLSM image of functionalized membrane with FITC antibody, (A) CLSM image of nitrocellulose nanofibrous membrane with FITC antibody functionalization, (B) CLSM image of nitrocellulose nanofibrous membrane without antibod y. Significant fluorescence emission at 530 nm verifies the antibody immobilization . 53 Figure 4.13 FITC antibody functionalized electrospun membrane: ( A) CLSM image verified that the f iber morphology retains after antibody functionalization. ( B) Fluorescence image confirmed that antibodies are attached on membrane after wash step by significant fluorescent emission at 435 nm. Figure 4.14 The SEM and optical microscope image of functionalized membrane with FITC antibody (A) SEM image of the electrospun nanofibrous membrane, (B) optical microscope image of nanofibrous membrane and silver electrodes after anti body functionalization. (A) (B) (A) (B) 1.0 µm 54 4.3.2 Aligned Nanofibrous Membrane by Parallel Electrode Electrospinning A special collector made of parallel electrodes was implemented in the NEU unit in order to create a membrane of highly aligned nanofibers [111] . In the vicinity of the parallel electrodes, t he electrical field l ine direction was split symmetrically into two trends pointing towards edges of the gap al ong the electrodes. The spun polymer jet followed the electrical field force , which stretched the synthesized nanofibers as being across the electrodes and orthogonal to the gap direction . The spinning process created a reproducible ultra -fine and highly aligned nanofibrous membrane, which wa s confirmed by SEM image ( Figure 4.15). Figure 4.15 SEM image of highly -aligned nanofibrous membrane synthesized by the p arallel electrode collector electrospinning . 55 A significant fluorescent emission was observed in the CLSM image at 600 × magnification , which verified strong antibody functionalization . Besides, the highly aligned fiber morphology remained intact after the functionalization process, which is shown in Figure 4.16 . Figure 4.16 The CLSM image at 600 ×: significa nt fluorescent emission verified strong antibody functionalization , and the highly -aligned fiber morphology still remained intact. 4.3.3 Biosensor Detection The process paramet ers and sensing condition s were determined to fulfil the biochemical conditions for the lateral flow immunoassay . The performance of lateral flow immuno biosensors depends on the capillary action which determines the effect of pathogen capture and impurity separation. The solution pH affects protein solubility, which affects antibody attachment and pathogen conjugation. Tris buffer was applied to sustain a neutral environment (pH 7.0) in the 56 assay to facilitate antibody -antigen reaction and reduce background noise [82] [112] [116] . Due to high surface area and low cross section of the MNPs and nanofibers , the electrospun MNP biosensor significantly improved the pathogen binding and impurity separation event. The applied antibodies for electrospun membranes was optimized to be 10 µg/in 2, which is more cost effective compared to conventional nitrocellulose planar material which typically requires 50 to 500 µg/in 2 following the same functionalization protocol. Figure 4.17 SEM image s of electrospun membrane after test . E. coli O157:H7 were effectively captured on the functionalized fiber mat . The pathogen capture by the immuno functionalized electrospun nanofibrous membrane was confirmed using SEM imaging . The l ateral flow test using E. coli O157:H7 sample were conducted on the electrospun membrane s with and without the immuno functionalization, respectively. After 57 flow equilibrium, the memb ranes were both washed and then dried for imaging . Figure 4.17 above demonstrated that the target pathogen can be effectively immobilized on the immuno functionalized membrane. Nonspecific material s were not observed in the image . Furthermore, the specifi city was based on affinity information of the purified and target -specific antibod ies as provided by the manufacture r. On the other hand , the untreated membrane failed to capture any organism ( Figure 4.18). Furthermore, the SEM image s verified that the unique nano structure of the nanofibrous membrane was retained after surface func tionalization and capillary flow action. Figure 4.18 SEM image s of electrospun membrane after test . No bacteria were observed in the nanofiber mat without functionalization . 58 The sensitivity and performance of the electrospun biosensor was verified by experiments using E. coli O157:H7 and BVDV of different concentrations , respectively . The real -time conductance signal of the electrospun MNP biosensor for samples of different E. coli O157:H7 concentrations are demonstrated in Figure 4.19 below. During the initial time around 50 sec after sample dispensing , the conductance signal increased with fluctuations as the antigen conductive MNP complex flowed to and along the capture pad. During the capillary action, t he immunoreactions tether ed the conductive MNPs in the nanofibrous membrane to form conducting bridges between the silver electrodes. After the flow reaches the absorption pad, the sensor conductance gradually decreased as the excess reagent were separated and absorbed by the capillary flow. When absorption achieved equilibrium after 8 min , the conductance signal became stable and was suitable for sensor reading for approximately 6 min . Afterwards the sensor conductance si gnal started to decrease rapidly due to drying effect, which caused reduction in the conductivity of polyaniline and antibody -antigen bonding [118] . As shown in Figure 4.19 below, t he biosensor conductance signal s of different test samples increased in proportion to the target concentration, and were all larger than that of the control sample. This result confirmed that a test sample with higher target concentration was able to create more electrical conducting bridges by the conductive MNP sandwich c omplex , which led to increase in biosensor conductance signal . The signals were differentiable at approximately 100 sec. We acquire d the result s at 10 min until the signal s became stable (with small variation s when the flow s approaching equilibrium ). 59 Figure 4.19 Biosensor conductance signal versus test time of E. coli O157:H7 samples with different target concentrations: aligned nanofibrous biosensor The pathogen detection results of the biosensor using unaligned nanofibrous membrane are illustrated in Figure 4.20. The biosensor demonstrated linear sensing response for test samples with different E. coli O157:H7 concentrations from 0 to 10 4 CFU/mL (Figure 4.21). Its sensitivity (detection limit) was measured to be 61 CFU/mL (P < 0.05, n = 3) . It could detect a contamination from a 1 mL sample without the optional magnetic concentration , since the infectious dose is thought to be less than 100 cells [119] . Th e detection limit is comparable with or better than other lateral flow biosensors reported in the recent literature (ranging from 101 CFU/mL to 104 CFU/mL ) [120] -[122] . Unpaired t-tests were used for two group comparisons and the associated P -value s were summarized in Table 4.2. 05010015020025030035000.02 0.04 0.06 0.08 0.1 0.12 0.14 Time (sec) Conductance (millisiemens) Control 10 CFU/mL 102 CFU/mL 103 CFU/mL 104 CFU/mL 105 CFU/mL 60 Figure 4.20 Biosensor (unaligned nanofib ers) conductance signal versus test time of E. c oli O157:H7 samples with different target concentrations: randomly oriented nanofibrous biosensor Figure 4.21 Biosensor test results (unaligned nanofib er) for E. coli O157:H7 demonstrate a linear sensing response from 0 to 10 4 CFU/mL. 50100150200250300350400450-0.02 00.02 0.04 0.06 0.08 0.1 0.12 0.14 Time (sec) Conductance (millisiemens) Control 10 CFU/mL 102 CFU/mL 103 CFU/mL 61 Table 4.2 P-value of biosensor test results (unaligned nanofiber) for E. coli O157:H7 Sample Pair (CFU/mL) P-value (n = 3) 0 (c ontrol ) vs 10 1 0.041 101 vs 10 2 0.025 102 vs 10 3 0.031 103 vs 10 4 0.023 The biosensor was also tested for potential application of virus detection . The results of BVDV test sample measurement s are illustrated in Figure 4.22. The associated P -value s were calculated and summarized in Table 4.3. The biosensor also exhibited linear response to different BVDV concentrations. The estimated viral concentration before dilution wa s 10 6 CCID/mL. The lowest detectable sample was 103 dilution (P < 0.05, n = 3) , which had virus concentration of approximately 103 CCID/mL and was equivalent to 1/1000 vir al concentration in the blood serum of infected bovine [117] . Figure 4.22 Biosensor test results (unaligned nanofibrous mat) for BVDV virus demonstrate the linear sensor response from 10 1 to 10 3 virus dilution and control sample. 62 Table 4.3 P-value of biosensor test results (unaligned nanofiber) for BVDV Sample Pair ( dilution ) P-value (n = 3) Control vs 10 3 0.003 103 vs 10 2 0.027 102 vs 10 1 0.029 The MNP biosensor was further improved by using aligned electrospun nanofibrous membrane, which was described in Chapter 4.2.1 . The test r esults for test sa mples of different E. coli O157:H7 concentrations are demonstrated in Figure 4.23 below . The associated P -values were calculated and summarized in Table 4.4. The aligned nanofiber biosensor exhibited a linear response in electrical resistance signal for E. coli O157:H7 concentration s from 10 1 up to 10 4 CFU/mL. The baseline signal appeared to be increas ing owing to the aligned nanofibers. However, no improvement in sensitivity was observed . The next level, 100 CFU/mL , was of extreme ly low concentration, which was difficult to detect due to variability in biochemi cal events and environmental noise . For concentrations at and above 10 5 CFU/ml, the measured resi stance of the biosensor became non dose -responsive. However, t he signals of 10 5 to 10 7 CFU/mL samples showed resistance that were still significantly below the control sample (0 CFU/mL ) (P < 0.05, n = 3) . This sensor behavior was due to the over -crowding effect due to the nature of sandwich immunoassays [123] . Above certain high concentrations, the binding site of the capture membrane was saturated by the excess amount of antigen, which kept the resistance signal from decreasing [124] . 63 Figure 4.23 Biosensor test results (aligned nanofibers) for E. coli O157:H7 demonstrate d a linear relationship between resistance signal and bacteria l concentration from 10 1 to 10 4 CFU/mL. The signal wa s significantly below the control from 10 5 to 10 7 CFU/mL. Table 4.4 P-value of biosensor test results (aligned nanofiber) for E. coli O157:H7 Sample Pair (CF U/mL) P-value (n = 3) 0 (control) vs 10 1 0.015 101 vs 10 2 0.004 102 vs 10 3 0.021 103 vs 10 4 0.040 104 vs. 10 5 0.963 105 vs. 10 6 0.484 106 vs. 10 7 0.766 64 The MNP biosensor fabric ated using electrospun membranes with/without fiber alignment both exhibited higher sensi tivity and wider linear response for microbial pathogen detection than that of conventional mesoporous membrane . To compare detection performance, biosensors using nit rocellulose porous membrane (Millipore, MA, USA) were fabricated following the same functionalization and sensor assembly process. Under the same testing , the electrospun biosensor remained linearly responsive from 101 to 10 4 CFU/mL, whereas nitrocellulose biosensor™s linear range was from 101 to 10 2 CFU/mL , as shown in Figure 4.24 below . The detection limit of the nitrocellulose biosensor was 10 2 CFU/mL (P < 0.05, n = 3). The associated P -values for nitrocellulose biosensor test results were calculated and summarized i n Table 4.5. Due to the unique nano porous structure and high surface area, the electrospun capture membrane provided more binding site and capillary flow action , which enhanced the immuno reaction and excess material separation. Compared to conventional mesoporous membranes, the contact area between the active mass transfer region of the nanofibrous membrane and the underlying substrate was greatly reduced. This substantially decreased electrical background noise in the direct charge measurement , which was caused by electronic charge interactions through different material interface . Hence, the elect rospun biosensor was able to detect not only the presence of lower target pathogen concentration , but also provide linear detection response. 65 Figure 4.24 Pathogen detection comparison of biosensors made of electrospun nanofibrous membrane and nitrocellulose mesoporous membrane . Table 4.5 P-value of biosensor test results (nitrocellulose mat ) for E. coli O157:H7 Sample Pair (CFU/mL) P-value (n = 3) 0 (control) vs 10 1 0.057 101 vs 10 2 0.038 102 vs 10 3 0.765 103 vs 10 4 0.023 66 4.4 Conclusions In this chapter, t he immun o-functionalization of PANI magnetic nanoparticles for biosensor application was successfully accomplished. The reaction solution and temperature were optimized for antibody functionalization affinity. The assay conditions were determined to be 0.25 mg of affinity purified mouse anti -E. coli O157:H7 monoclonal antibody for 2.5 mg PANI MNPs in a total volume of 250 µL PBS solution [115] . The immuno PANI MNPs were successfully implemented for magnetic separation for impurity removal and sample concentration. To fully investigate E. coli O157:H7 detection, the PANI MNP based immunoassay was successfully integrated into an elec trospun lateral flow biosensor system . Due to the unique nanostructure and biocompatibility of PANI MNPs , the biosensor had linear detection response of E. coli O157:H7 sample concentration from 10 1 to 10 4 CFU/ mL and exhibited a detection limit of 61 CFU/mL. The application of the biosensor can be extended to other microbial or viral organisms by appropriately changing the antibodies . Due to the unique property of the magnetic nanoparticles , this simple measurement system makes applications possible for low cost detection and rapid on field testing. The MNP based electrospun lateral flow biosensor demonstrated excellent detection performance and rapid response compared to other conventional E. coli detection methods. The total det ection time require d 15 min, which includes 5 min of magnetic separation and 10 min of lateral flow process (until the flow s are stable) and data acquisition. 67 To further improve detection time, a new type of advanced biosensor was designed based on NMR and MNPs in our lab . It s develo pment , optimization and experimental testing are summarized in Chapter 5 . 68 This chapter is adapted from our recently published work in the journals of Biosensors and Bioelectronics , and IEEE Transactions on Nanotech nology : Yilun Luo, Steven Nartker, Hanna Miller, David Hochhalter, Michael Wiederoder, Sara Wiederoder, Emma Setterington, Lawrence T. Drzal, Evangelyn C. Alocilja. Surface functionalization of electrospun nanofibers for detecting E. coli O157:H7 and BVDV cells in a direct -charge transfer biosensor. Biosensors and Bioelectronics. 201 0. 26(4):1612 -1617 . DOI: 10.1016/j.bios.2010.08.028 Yilun Luo, Steven Nartker, Michael Wiederoder, Hanna Miller, David Hochhalter, Lawrence T. Drzal, E vangelyn C. Alocilja. Novel Biosensor Based on Electrospun Nanofiber and Magnetic Nanoparticles for the Detection of E. coli O157:H7. IEEE Transactions on Nanotech. 2011. 11(4): 676 Œ 681. DOI: 10.1109/TNANO.2011.2174801 69 CHAPTER 5 BIOSENSOR BASED ON NMR AND MAGNETIC NANOPARTICLES FOR PATHOGEN DETECTION 5.1 Introduction The Nuclear Magnetic Resonance (NMR) is a versatile scientific instrument to analyze the nuclear magnetic properties of atomic nuclei and determine chemical properties of biomole cules . As its signal is powerful to penetrate turbid raw sample s, NMR instruments have wide applications in non-destructive biomedical diagnosis, which can simplify sample preparation process, and save analysis time [125] . However, the commercial NMR systems are expensive, bulky and heavy (around 1,000 kg), which limit their application for portable pathogen detection [126] . A portable NMR system was developed using integrated circuit technique [133] . It was designed to detect avidin and canc er cells by measuring the test sample™s NMR relaxation, which was proportional to the immunological clustering of magnetic nanoparticles [134] [140] . The fabrication for integrated circuit is high in cost. And besides integrated circuits, it sti ll required additional systems, such as signal generator , filters, and data acquisition system, which considerably increase the system cost and size. The ir sensitivity for cancer cell detection was in the range of 1000 CFU/mL. The detection of E. coli O15 7:H7 is time consuming and requires complex instruments and extensive training. The Sorbitol -MacConkey (SMAC) agar method is used for identification but takes about 2 to 4 days, including culture, morphological identification, and confirmation techniques [145] [146] . The polymerase chain reaction (PCR) based detection assays are sensitive and accurate [147] . However, the PCR techniques require complex sample preparation steps, such 70 as DNA extraction and amplification, which increase addition al diagnosis time. Fast detection methods have been reported based on immunological detection. However, most of these diagnosis systems have sensitivity limit greater than 100 CFU/mL, detection time of 1 hour, and are not applicable for in-field applicatio ns [148] [149] . To minimize the spread of infection and costly product recall, a rapid, sensitive, and portable detection of E. coli O157:H7 is essential in food supply and healthcare applications. This chapter reports a new design of portable NMR system including NMR probe and transceiver for a biosensor to detect Escherichia coli O157:H7 using the magnetic nanopa rticles as biomarker. The system is low in cost and small in size, and it is thus suitable for portable pathogen detections. 5.2 Materials and Methods 5.2.1 Design of Portable NMR Biosensor A complete NMR system was designed consisting of proton NMR probe, high pow er and high sensitivity trans mitter and re ceiver, field -programmable gate array (FPGA ) based pulse controller , and communication interface. The system utilized a n ultra -strong permanent magnet , PM-1055, by Metrolab Instruments Inc . (Geneva , Switzerland ). Its magnetic field strength is 0.49 Tesla with field homogeneity of 10 ppm. It has a compact size of 80 mm diameter and 55 mm height , and weigh s 1250 g [150] . A gauss meter was used to calibrate the magnetic field strength and determine d its most homogeneous region to place the test sample for NMR detection . A solenoid coil of 5 mm diameter and 5 mm length was fabricated in lab . A matching network using low -paras itic capacitor trimmers was designed to achieve impedance matching and high -quality factor, Q [151] . The signal to noise ratio (SNR) was optimized in the NMR design as described in the following sections . 71 The portable NMR system was designed and built in the Nano -Biosensor s Lab. It consist ed of an NMR transmitter, a transmit/receive (T/R ) switch, an NMR probe, and an NMR receiver , which are illustrated in Figure 5.1. In t he NMR transmitter , the FPGA control s a Direct Digital Synthesis (DDS) to generate s a 19.9 18 MHz RF signal for NMR excitation and signal demodulation in the receiver. After noise removal by a low pass filter (LPF) , the RF signal was amplified by a linear power amplifier to reach 20 W, capable of excit ing water nuclei resonance spin inside the NMR probe. The input to the power amplifier was switched off to reduce its background electrical noise. The T/R switch was design ed using high speed crossed diode s and quarter wavelength transmiss ion line s to protect a low noise amplifier from high -power during excitation and block the electrical noise from the NMR transmitter during receiving. In t he NMR receiver , the weak NMR signal was amplified by a low noise amplifier , which has a high amplification gain and very low noise. It is able to detect signal in 0.1 µV , which is reemitted from the excited nuclei spinning in resonance. After noise removal using band pass filter s (BPF s), the NMR signal was demodulated using RF mixer s and LPF s. Quadrature demodulation was implemented to obtain the direct and quadrature components of the NMR signal to enhancement signal . With excellent concurrent calculation capability and high integration, an embedded system was designed in the FPGA using a mul tilayer state machine to receive commands , display results via HyperTerminal graphical interface , control NMR transmitter gain and frequency , acquire data through an analog to digital converter (ADC), data processing, and provide precise control to generate versatile NMR pulse sequence. 72 Figure 5.1 System Architecture of the Portable NMR , including FPGA control, modulation and demodulation system, and NMR sub -systems. A prototype of the portable NMR system has been built as shown in Figure 5.2. The magnet holder, NMR probe holder , and gauss meter probe holder were designed using aluminum and wood , which does not interfere with the magnet ic field . An X -Y-Z precision linear positioner was used to determine the most homogeneous region of the magnet , and adjust the NMR sample position, to optimize the NMR sensitivity . The prototype system was built using a low -cost amplifier , 50B power amplifier (Henry Radio), and AU -1467 linear amplifie r (Miteq Inc.). The overall size of the prototype was 32 mm × 24 mm × 14mm. The system can be integrated in a mini personal computer enclosure , 20 mm × 18 mm × 8mm for better portability and electromagnetic compatibility (EMC) performance. A commercially a vailable low-cost 5 mm NMR tube was used as test sample holder and is reusable by acetone washing. The sample volume of the prototype system was could be further reduced by using thinner NMR tubes , such as a standard 20 µL NMR tube. NMR Probe Crossed Diode Splitter LPF LPF DDSSIN COS Crossed Diode Low Noise Amplifier Finite State Machine 1/4 Wavelength Transformer Matching Networks BPFMixer Mixer LPF LPF ADC DF-3dBFrequency Control Linear Power Amplifier Pulse Control Communication FPGA Deblank Control FFT73 Figure 5.2 System prototype of the portable NMR , including a palm -sized permanent magnet, NMR transmitter, NMR receiver, T/R switch , sample holder, and NMR probe. 5.2.2 Design of Portable NMR Duplexer The NMR duplexer is an RF switch in the NMR system , which automatically switch the NMR probe to connect with the power amplifier (PA) of the NMR transmitter or the low noise amplifier (LNA) of the NMR receiver. It was design ed and built using a crossed diodes pair (1) in series with a quarter -wave impedance transformer follow ed by a second crossed diodes pair (2) 74 connecting to the ground , as shown in Figure 5.3. A second quarter -wave impedance transformer and crossed diodes pair (3) were added t o improve the blocking of high -power signals . Figure 5.3 Complete set up of NMR coil antenna and duplexer switch for the NMR transmitter and receiver During the NMR excitation process , the duplexer transmits the 20 W high -power RF signal from the power amplifier to the NMR probe while blocking that from the LNA to prevent any over rating damages. The antiparal lel crossed diode pair operates as a high -frequency switch for RF signals . It is open for signals larger than the dio de™s built -in potential and is closed for small signals (Figure 5.4). The 20 W RF signal is greater than the built -in voltage of the diodes . The crossed diodes pair (1) is turned on as both of its diodes have been given an external forward bias. Transmission Line Transformer Crossed Diode 2 NMR Receiver NMR Transmitter Crossed Diodes 1 Transmission Line Transformer Crossed Diode 3 NMR Antenna L0R075 Figure 5.4 The crossed diodes pair shows high impedance for low voltage s, and low impedance for high votlages . The threshould is the built -in voltage. For the quarter -wave impedance transformer, the relationship between its input impedance, Zin, and its load impedance , Zload , is as defined in Equation 2. = Equation 2 where Z0 is the characteristic impedance of the transmission line, 50 . The high -power RF signal after the conducting crossed diodes pair (1) is sufficient to turn on the crossed diodes pair (2) , which makes Zload equal to the forward re sistance of the diode (close to 0 ). This make s Zin becom e very high as =/ . The more sets of quarter -wave impedance transformer and crossed diodes pair are added, the higher the impedance at the input port . As a result, the NMR duplexer circuit makes the NMR probe being close d-circuit to the NMR transmitter and disconnected from the NMR receiver (Figure 5.5). Resistance Voltage across the diodes pair +Vbuilt-in -Vbuilt-in Crossed diodes pair 76 Figure 5.5 NMR coil antenna and duplexer switch based on quarter -wave impedance transformer : transmitter ON During the NMR signal pick up process , the power amplifier is of f. However, its background noise is still overwhelming for the LNA, whose detection range is in 0.1 As t he crossed dio des pair (1) is reversed biased , it becomes open -circuit to block the signal noise from the power amplifier. Meanwhile, under NMR echo the crossed diodes pair ( 2) is still reversed -biased and therefore is open -circuit from the ground , which makes Zload equal to the input impedance of the LNA, 50 . This makes Zin becom e 50 as =/ , which is equivalent to be closed -circuit on its input port. As a result, the NMR duplexer circuit makes the NMR probe being open -circuit to the NMR transmitter and close d-circuit to the NMR receiver ( Figure 5.6). Transmission Line Transformer Crossed Diode 2 NMR Receiver NMR Transmitter Transmission Line Transformer Crossed Diode 3 NMR Antenna L0R077 Figure 5.6 NMR coil antenna and duplexer switch based on quarter wavelength transformer: transmitter OFF The quarter -wave impedance transformer is built in the Nano -Biosensor s Lab. The line length is calculated , ¼ = 8.1515™™ , for the 19.918 MHz NMR frequency. Different diodes are tested to optimize the NMR duplexer performance. It was observed that PIN (p-type/intrinsic/n -type) diodes with fast switching and low patristic internal resistance yields the best performance. Finall y, one set of quarter -wave impedance transformer gave a -20.14 dB signal reduction from the input to the output . Hence, an additional quarter -wave impedance transformer was added to the receiver end of the duplexer for improving filtering of the source sig nal to -33.98 dB in total . This is sufficient for blocking the background noise from the power amplifier so the weak NMR signal could be detected by the LNA . 5.2.3 Design of Portable NMR Antenna The NMR probe is an RF antenna, which is installed on the NMR tube to deliver high -power RF energy to sample for NMR excitation and receive echo signal from NMR nucl ei precession . The pNMR probe was designed consisting of a coil antenna and the matching networks to optimize both functionalitie s. NMR Receiver NMR Transmitter Crossed Diode 1 NMR Antenna L0R078 5.2.3.1 Antenna Frequency The operating frequency of the pNMR probe is determined by the Larmor or precessional frequency, which defines the precession rate of the magnetic moment of the proton under an external magnetic field. The Larmor frequency of nuclei of a substance can be calculated from the Larmor equation as shown in Equation 3 [154] . = Equation 3 where is the Larmor frequency in MHz, is the gyromagnetic ratio in MHz/Tesla, and B0 is static magnetic field strength in Tesla. The of proton, H1, equals to 42.58 [155] . Hence the frequency of pNMR probe is calculated as 19.918 MHz. 5.2.3.2 Signal to Noise Ratio NMR signal is radio frequency radiation released from excited spin state to thermal equilibrium by a small number of hydrogen nucle i in the NMR antenna coil . Due to the weak magnetic field by the handheld magnet, t he NMR signal is small , typica lly in the range of 0.1 , and hence difficult to detect . Noise is produced from multiple sources in the system, such as the coil antenna , discrete circuit component s, cables , the amplifier, and the receiver. Thus , the signal to noise ratio , SNR, was studied in the design of the portable NMR system to achieve biosensor detection from the complex superimposed background nois e. The SNR of a n NMR system can be calculated as the voltage level of the received NMR signal divided by that of the noise signal . During the NMR reception, the voltage difference in the NMR 79 probe , or the electromotive force (EMF) , is induced by a time -varying magnetic field flux passing through a surface, S enclosed by path C, which can be calculated by the Faraday™s law o f induction as in Equation 4. = = Equation 4 where is the electric field , is the magnetic field , ,and are the close integral and surface integral enclosed by path C, respectively . Based on the principle of reciprocity, for the single coil used in NMR for both transmission and detection, the receive field can be assumed to be equal to the transmit field [156] . Hence, t he voltage signal induced by magnetic dipole, , can be calculated using the transmit field strength , , at the same location while current is applied to the same coil , C, during the transmission (Equation 5). = Equation 5 The value s of EMF in all locations of the sample are integrated to calculate the signal produced by the whole sample (Equation 6). = Equation 6 where Vs is the sample volume in the RF coil , and is the magnetization along the x -y plane (perpendicular to the field plane of the permanent magnet , z). 80 The value of B1 is determined by the coil shape. For the long -cylind er shaped coil antenna selected in this pNMR designed , it can be considered as homogeneous over the sample sensing volume . Hence, the Equation 6 can be simplified to as Equation 7. = cos Equation 7 where K is an inhomogeneous factor, is the Larmor equation as defined in Equation 3, is B1™s component along the x -y plane. The total nuclear magnetism, can be calculated as shown in Equation 8. =()(+1)/3 Equation 8 where N is the spin density at resonance per unit volume, unit volume, is the gyro magnetic ratio, I is the main total angular momentum quantum number (I = 0, ½, 1, 3/2, 2, – ), and TS, is the sample temperature. For water proton, a spin ½ system, Equation 9 can be derived using Equation 8 and the population difference between ½ and ½, and further simplified to obtain Equation 10. = =( ½½)/2 Equation 9 =1.0210 /4 Equation 10 81 Therefore, the signal to noise ratio, SNR , can be calculated as Equation 12 by substituting Equation 10 to Equation 7. = Equation 11 =1.0210 4 Equation 12 where is the rms voltage of noise at the output of the NMR coil , and is the magnetic field in the transverse plane produced by the unit current in the NMR coil . As a result, to detect the weak NMR signal from the background nois e, the SNR should be maximized in the NMR design by optimizing the antenna design for field homogeneity and transmi ssion efficiency , by increasing sample volume and transmit power, and by reduc ing noise to the receiving amplifier. 5.2.3.3 Antenna Quality Factor The quality factor, Q could improve SNR by reducing noise and improv ing antenna efficiency for nuclei excitation . It is a concise performance metric for bandwidth of an antenna [157] . It is defined as the ratio of the power stored in the reactive field over the radiated power [158] . High Q antennas store a lot of power in the near field, which is desirable for NMR proton excitation . In addition, a high value of Q also corresponds to a narrower bandwidth. The higher the quality factor of the NMR probe, the better the signal -to-noise ratio. Hence, high quality factor is desirable for the NMR receiver to discriminate the NMR signal against noise of different frequencies other than the Larmor frequency ( Figure 5.7). 82 Figure 5.7 A high Q NMR antenna corresponds to higher selectivity, and is more immune to noise for the NMR receiver [153] The coil antenna is modeled as an ind uctor in series with parasitic resistance as shown as blue components in Figure 5.8. As a L -R circuit, the quality factor of a coil antenna is calculated as Equation 13. = Equation 13 where is the frequency, L0 is the inductance of the coil antenna, and R0 is the parasitic resistance of the coil antenna. 5.2.3.4 Sample Volume Based on Equation 12, the SNR is linearly proportional to the sample volume , which is determined by the coil antenna. NMR antennas in other portable NMR systems , such as flat solenoid and sp iral 83 surface coil , are small in size , but are difficult to achieve uniform field and high volume in the sample due to the ir coil shape s [132] [134] . 5.2.3.5 NMR Antenna The antenna of the pNMR system selected the design of long -cylinder shaped coil, which is beneficial to achieve high field homogeneity and sample volume . It was fabricated using high -quality enameled copper wire by winding on a steel rod of the same diameter as the NMR sample for L0 and R0 using an Agilent LCR meter at 19.918 MHz , to optimize quality factors and sample volume for NMR performance. The test results indicate that coil by 8 turns mm wire (8 mm in length and 5 mm in inner diameter ) is the optimal for NMR antenna . Antenna made with thinner wires leads to higher parasit ic resistance , while thicker wires end up with fewer turns , thereby reduc ing inductance. Hence, b oth cases have a negative impact on the quality factor. Such design s lead to a big sample with volume of 18 , which is more than 10 times larger compared to other portable NMR systems reported in the literature [132] [134] . 5.2.4 Design of Matching Networks of the pNMR Probe The function of NMR probe is to deliver RF energy from the power amplifier on the transmitter and pick up NMR echo signal to the low noise amplifier on the receiver. Impedance matching is an essential design practice to maximize the power transfer and minimize signal reflection from the load. 84 Based on the transmission principle of wave or RF signal , the reflection is calculated by reflection coefficient , the complex ra tio of the voltage of the reflected wave to that of the incident wave Equation 14. = =+ Equation 14 where is the reflection coefficient from load to source, VReflect and VIncident are the voltage of the reflected and incident wave, the ZS is the source impedance, and ZL is the load impedance . Thus, when the input impedance of the electrical load matches with the output impedance of its corresponding signal source , the reflection equals to zero , therefore transmitting all the RF signal to the load. The coil antenna is an inductor but with para sitic resistance ( Figure 5.8). Hence, a matching network is required to transform its impedance to 50 in order to match with the output impedance of the power amplifie r and the input impedance of the low noise amplifier (both being 50 ) for maximum signal/ power transmitting and receiving . 85 NMR Antenna L0CpCsR0ZpZin Figure 5.8 The NMR probe, consisting of NMR coil antenna and a matching network, was designed to achieve high quality factor to optimize NMR signal acquisition. The matching network is designed based on series -parallel resonance circuit compri sing of a parallel capacitor, Cp, and a series capacitor, Cs, onto the coil antenna. The total impedance can be modeled as, Equation 15, where Zin is the input impedance of the NMR probe, including the NMR coil antenna and the matching network. =1+1(+)1++ Equation 15 The complex impedance of Zin can be decomposed as the real component, (), and the reactive component, (), as Equation 16. () and () can be further calculated based on the resistance, capacitance, and inductance of each circuit component. 86 =()+() Equation 16 ()=(1)+ Equation 17 ()=1+11+ Equation 18 As indicated in Equation 17 and Equation 18, the real part of Zin can be adjusted by Cp to match with that of the power amplifier output , ZPA. Meanwhile, the inductive reactance of Zin can be resonated out by Cs, therefore corresponding to Zin = ZLNA = ZPA = 50 . The resistance and inductance of the coil antenna was measured as described above to determine the capacitance of Cs and Cp. The coil antenna, that with an empty NMR tube, and with DI water were measured by an Agilent LCR meter, respectively. The measurement and calculation results are summarized in Table 5.1. Table 5.1 Matching network design based on NMR probe characterization through resistance and inductance measurement R0 () L0 (nH) Cp (pF) Cs (pF) Coil only 1.36 685.70 77.8 15.43 NMR tube 1.45 712.29 74.41 15.33 NMR tube with DI water 1.46 714.69 74.12 15.32 87 It is observed that inserting the NMR tube impact s both resistance and inductance of the coil antenna, while that of water , the NMR target, was negligible. The NMR tube is of instrumental grade and commercially available. It is made of pure high -quality quartz and is non -mag netic, which will not affect the electromagnetic field generated by the coil antenna . Hence, i ts antenna impact is mainly by coil enlarg ement due to the insertion. Considering this factor , the Cp and Cs are accurately calculated as 74.12 pF and 15.32 pF, respectively. A 50 pF fixed capacitor in parallel with a 50 pF variable capacitor were chosen for Cp, while a 10 pF fixed capacitor in parallel with a 10 pF variable capacitor were chosen for Cs. It is worth noting that the quality of the capacitor: high Q and low parasitic resistance , is essential to achieve high quality factor for the NMR probe . Based on the d esign, a n NMR probe comprising of NMR coil antenna, NMR sample tube, tuning circuit, and non -magnetic probe holder is fabricated in the Nano -Biosensor s Lab. Its RF performance was tuned and evaluated as described as follows. 88 Figure 5.9 NMR probe compris ing of NMR coil antenna, NMR sample tube, tuning circuit, and non-magnetic probe holder is designed and fabricated in the Nano -Biosensor s Lab. 5.2.5 Evaluation of the Portable NMR Probe and the Matching Networks The performance of the NMR probe , including reflection coefficient ( S11) and quality factor ( Q), were evaluated using a Keysight E5062 vector network analyzer (VNA) . During the testing, the NMR probe system , including the coil antenna and the matching networks , is connected to the port 1 of the VNA , as shown in Figure 5.10. The VNA transmits RF signal s to the NMR probe at frequencies scanning from 0 to 40 MHz and measure s the reflected power from the probe for each frequency point . The VNA measures the reflection coefficient, S11, which describes how much of a RF signal is reflected by an impedance discontinuity in the transmission system . It is defined by the two -port theory as Equation 19 [152] . NMR Antenna NMR Sample Tube Tuning Circuit 89 =20 =10 Equation 19 where Er and Ei are the reflected and incident fields , Pr and Pi are the reflected and incident power , respectively. The S11 measures how much power is reflected back to the power amplifier . Since the input power to the probe equals to the sum of the incident power and the reflect ed power, the S11 measures the probe™s total efficiency of power transmission into the sample for the NMR excitation. Figure 5.10 The NMR probe evaluation using a Keysight E5062 VNA for reflection coefficient and quality factor . It s matching network was tuned while measuring using the VNA to optimize the reflection coefficient (S11) and achieved -23.6 dB. Preflect Pincident Keysight E5062 VNA NMR Probe & Matching Networks 90 The matching networks was tuned by adjusting two tuning capacitors, Cp and Cs to minimize the reflection coefficient at the NMR operating frequency, 19.918 MHz . A S11 of -23.6 dB was achieved for 19.918 MHz , which indicate d that 0.4% of input power was reflected and 99.6% was transmitted for NMR (Figure 5.11). Figure 5.11 The NMR probe was designed to achieve high quality factor to optimize NMR signal acquisition. It s matching networks was tuned to optimize the reflection coefficient using a VNA and achieved -23.6 dB at 19.918 M Hz. The matching networks consist ed of two capacitors, which were reactive circuit elements. The remaining signal that was not reflected was assumed to be transmitted and passed to the sample. 91 Thus, t he quality factor of the NM R probe could be calculated from the reflection coefficient results using Equation 20 below . = Equation 20 where f0 is the antenna frequency , and is the antenna bandwidth . Based on the S11 experime ntal data (Figure 5.11), the bandwidth is the frequency difference at 3 dB points below the peak amplitude , where f0 = 19.918 MHz at -23.6 dB, f1 = 19.898 MHz at -20.6 dB, and f2 = 9.927 MHz at -20.6 dB. Hence, the quality factor could be calculated as, Q = 19.918 / (19.927 - 19.898) = 686.8. However, an accurate measurement of the Q-factor using a one-port measurement is difficult, and two -port techniques should be implemented [159] . To compar e performance , two NMR probe s (from Bruker Inc. ) were evaluated for the quality factor using the VNA as described above. After tuning the matching ne tworks, t heir Q™s were measured as, 64.04 and 538.32, respectively. Liu et al . built a solenoid al coil -based NMR probe with Q of 200 [132] . Sun et al . built a planar microcoil -based NMR probe with Q of 16 [134] . In summary, the pNMR™s probe , designed and fabricated in the Nano -Biosensor s Lab, achieved better Q than those in the commercially available system s and in the existing literature . Hence, the probe of the pNMR is very efficient at delivering energy to the sample and achieving a high SNR for the detected NMR signal , which are beneficial for detect ing lower concentration pathogen s in complex sample s. 92 5.2.6 NMR Power Amplifier The power amplifier is used in the pNMR system to amplify the RF pulses to a specific power level in order to rotate the sample magnetization to the desired angle effectively and repeatably. When the input impedance of the NMR probe is tuned to match with the power amplifier output (50 ), all the output power of the amplifier will flow in the coil without any reflection. The amplifier power required for magnetization rotation can be calculated by the equivalent resistance, Reff, and the NMR probe current , I (Equation 21). =12 Equation 21 = + + Equation 22 The equivalent resistance is contributed by the sample, , the coil, and matching network circuit, , and radiation resistance, (Equation 22). This was measured on the NMR probe loaded with a test sample using a VNA at the NMR frequency of 19.918 MHz . The measurement result is shown in Table 5.1. The probe current is determined by the induced magnetic field strength , , the pulse width, , and the flip angle as shown in Equation 23. One example of the application is the Carr -Purcell -Meiboom -Gill (CPMG) experiment, which is based on RF pulses to achieve . 93 ==( ) Equation 23 The magnitude of the induced magnetic field is determined by the antenna . For the coil antenna of the pNMR, it can be modeled as Equation 24. =1+ Equation 24 where l is the coil length, d is the coil diameter, and n is solenoid turn number of the coil . Hence , the coil current , I, can be obtained from Equation 23 and Equation 24. Substituting I to Equation 21, the required amplifier output power can be calculated as defined in Equation 25. =12 (+) Equation 25 To generate a 90 magnetization RF pulse in 7 or a 180 RF pulse in 14 , the transmit power was calculated to be, 13.57 W using Equation 25. The required amplifier output power was calculated to be 2 0 W, taking into account of the cross ed diode loss and the harmonics loss , which were measured as 1.11 dB and 0.76 dB , respectively . These loss es occurred in the T/R switch and power amplifier , which were before the NMR probe . Therefore , the y did not affect the NMR probe evaluation by Q measurement in section 5.2.5. NMR detection relies on sinusoidal RF wave s to excite the nuclear spin s of the test sample. Therefore, a linear power amplifier is required for this application. In addition , the NMR signal is 94 weak, ranging around 0.1 The NMR reception can still be interfered by the amplifier noise leak ing to the receiver, even after the duplexer s has swi tched the antenna from the power amplifier to the receiver . To solve this problem, a high -speed solid -state RF switch, ZYSW -2-50DR ( mini -circuits Inc.) was selected to switch the RF signal from the input of the power amplifier to a 50 RF load [135] . The input switching along with the CPMG sequence puls ation were precisely controlled by an embed ded software programmed in Very High Speed Integrated Circuit H ardware Description Language (VHDL) on the Spartan -3 FPGA control board running at 300 MHz operation speed [136] . As a result , a linear amplifier with continuous wav e (CW) capability was required to fulfill the requirements of transient response and high -speed switching . Based on the pNMR requirement s on frequency, power, and operati ng mode, w e identified the 50B HF power amplifier from Henry Radio Inc. as suitable for the pNMR application . The RF performance of the 50B HF amplifier is summarized in Table 5.2 below [137] . Table 5.2 RF performance of Henry Radio 50B HF power amplifier , including output power, frequency range , and operating mode . Henry Radio 50 B HF Power Amplifier Parameter Specification Watts In/Out CW Œ 1 mW to 40 W Bandwidth 1.8 to 30 MHz Mode SSB, CW or FM DC Volts/amps 26.0 VDC - 4 A 95 To transfer 20 W output power to the NMR probe, the input power was calibrated through testing of the NMR transmitter , consisting of the RF switch, power amplifier, crossed diodes, and wavelength duplexer s, using the testing set up shown in Figure 5.12. A Tektronix AFG3021 function generator was connected to the transmitter input to provide RF signal at 19.918 MHz. An oscilloscope was connected to the transmitter output to measure the RF si gnal on a 50 load simulating the antenna , and the output™s sinusoidal waveform , switching performance, frequency , and RF power were verified and measured. Transmission Line Transformer Crossed Diode 2 Power Amplifier Crossed Diodes 1 Transmission Line Transformer Crossed Diode 3 50 Ohm 50 Ohm 50 Ohm Figure 5.12 Testing set up of the NMR transmitter using the function generator, 50 loads and oscilloscope to calibrate the input power required for pNMR application . The input power of the NMR transmitter was calibrated to be 0.02 mW, 0.06 mW, and 0.0 8 mW to achieve output power s of 5.2 W, 16.0 W, and 20.0 W (45.6 Vpp , 80.0 Vpp , and 89.5 Vpp ), as illustrated in Figure 5.13. The amplification gains were measured as 55.0 dB, 54.5 dB, and 5 4.0 dB, respectively. It is specified to be 53.0 dB by the supplier. Hence, the calibration experiment is essential to make an accurate detection for NMR. NMR Transmitter Output 96 Figure 5.13 The input power to the NMR transmitter was calibrated to achieve 20 W output to the NMR probe . 5.2.7 NMR Low Noise Amplifier To detect NMR signal of 0.1 V in a noisy environment , the receiving amplifier is required to be low in its own noise and have high a gain higher than 60 dB to allow RF demodulat ion by the mixer . We identified AU-1467 low noise amplifier (LNA) from Narda Miteq Inc. with an amplification gain of 67.9 dB and noise figure of 1. 2 dB suitable for this NMR application [138] . Its RF performance is summarized in Table 5.3 below. 97 Table 5.3 RF performance of Miteq AU-1467 low noise amplifier , including frequency range, gain, and noise figure . Miteq AU -1467 Low Noise A mplifier Parameter Specification Frequency Range 10-600 MHz Gain 65 dB Min, 67 dB Typical Gain Flatness +/- 1 dB Max Input VSWR 2.0:1 Max Output VSWR 2.0:1 Max Noise Figure 1.2 dB Output P1dB +12 dBm Min The LNA was tested using the NMR receiver setup ( including transmission line transformers ) as shown in Figure 5.14. A sinusoidal test signal of 0.1 V amplitude and 19.918 MHz frequency was generated by the function generator . It was attenuated to be 0.1 V using Mini -circuits RF attenuator s (120 dB total a ttenuation) to simulate the NMR output signal . Finally, the attenuated signa l was transmitted to the LNA through two s so that the circui try remained the same as it was in the full pNMR system . 98 Transmission Line Transformer Crossed Diode LNA Transmission Line Transformer Crossed Diode Attenuator Figure 5.14 Testing set up of the NMR receiver using the function generator, attenuators , and the oscilloscope to evaluate the LNA with . The LNA output was evaluated using a Tektronix oscilloscope. With the scope™s built -in amplification by 40 dB, the amplitude of the LNA output was measured to be 24.8 mV, indicating that the amplification gain of LNA is 67.9 dB (fulfill the design requirement). The out put waveform was sinusoidal with frequency measured to be 19.918 MHz. These test results demonstrated that the LNA could A previous attempt was to use a Sonoma Instrument 310 amplifier as the LNA of the pNMR [139] . Its performance of bandwidth and noise figure both meet the system requirement (Table 5.4), however its gain of 32.5 dB was insufficient to recover the weak NMR signal. It could not detect in the set up above or the pNMR detection of pure water sample . 99 Table 5.4 RF performance of Sonoma Instrument 310 low noise amplifier , including frequency range, gain, and noise figure. Sonoma Instrument 310 Low Noise A mplifier Parameter Specification Frequency Range 9 kHz - 1 GHz Gain 32.5 ± 1.5 dB Gain Flatness +/- 0.5 dB Max Noise Figure 1.8 dB typ. Output P1dB +10 dBm 5.2.8 NMR Detection using Magnetic Nanoparticle A nuclear magnetic resonance (NMR) instrument allows analyzing the content of a sample and its molecular structure by measuring frequency and duration of electromagnetic signal emitted from nuclear spin relaxatio n [162] . Both features of the NMR signal are det ermined by the atomic and molecular properties of the sample as well as its applied external magnetic field. To better analyze the sample™s intrinsic properties, the magnetic field needs to be close to the ideal states : being strong in strength in order to achieve dispersion of response frequencies, while of very high homogeneity and stability over the entire sample space in order to deliver frequency resolution to reveal the details of chemical shifts and the Zeeman effect. To fulfill these conditions, the NMR system has complex and expensive magnet system to achieve highly uniform magnetic field ranging from 1.5 to 20 Tesla, which needs to be made of rare -earth stron g magnetic magnets with complex design for field uniformity, or even by liquid - helium/nitrogen cooled superconducting 100 coils consuming large currents. Although the configuration produced accurate results, the NMR instruments were rather large, heavy, not p ortable, and expensive [163] [164] . This has made the NMR technique less pr actical and less useful for on -site applications , as demonstrated in Figure 5.15. Figure 5.15 The Conventional nuclear magnetic resonance (NMR) spectrometer , Bruker 700 MHz NMR system [165] . Magnetic nanoparticle (MNP) is a unique material with a range of desirable properties, including nano sizing, high surface -to-volume ratio, self -prevention from agglomeration, and strong magnetizat ion strength (hundreds of emu) [166] . Each MNP can function as a separate nanomagnet. Owing to its high quantity -to-volume ratio, a tiny amount of MNPs could achieve 101 tens of millions of nano magnets, each interacting with the external magnetic field, and effectively inducing field distortion. As aforementioned in the previous paragraph, since nuclear spin interaction is very sensitive to the uniformity of the external magnetic field, when a few MNPs bound to their intended molecular target through antibody affinity, this can lead to an effective reduction in the bulk spin -spin relaxation time ( T2) of the surrounding water molecules. Using a palm -sized permanent magnet with lower strength and uniformity, the difference in T2 shortening caused by different target concentration could be effectively detected without the need for an instrument -level NMR system . In addition, owning to the excellent penetrability of the detection signal, MNP -based NMR biosensor can be used to mea sure turbid samples with less sample preparation. The detection time is shorter than those with surface -based techniques relying on the targets ™ molecular diffusion to the sensing elements. The measurement is non -destructive, allowing characterization by o ther detection methods afterwards. These advantages make the MNP -NMR ideal for fast, simple , and high -throughput sensing applications, especially in portable form factor. Being more flexible and smaller in size, the MNP -NMR is suited for on -site and field -based measurements. This class of biosensors can be developed and manufactured in low cost and are less costly to operate and maintain as compared to bulky conventional NMR instruments. In the Nano -Biosensor s Lab, we developed and built a portable MNP -NMR biosensor using a palm -sized magnet of 0.5 Tesla field strength [160] [161] . We also developed an antibody -functionalized superparamagnetic MNP for the biosensor application [160] [167] . Each superparamagnetic -Fe2O3) magnetic core and a polyaniline shell. 102 The synthetization process was tuned for size and magnetic property, achieving a uniform average diameter of 80 nm and a total distribution of 50 nm to 100 nm. The MNP has strong saturation magnetization (M S), 38 emu/g as indicated by M -H hysteresis measur ement. The coercivity (H C) was found to be the same as the unmodified Fe 2O3 particles, 180 Oe. This indicates that the polyaniline shell did not affect the MNP™s anisotropy energy, which is essential to modify both the NMR precession frequencies and the th ermodynamic occupation probability of the crystal magnetic states. Hence, our superparamagnetic MNP is favorable to shorten the NMR relaxation, and to be used as proximity sensors to amplify molecular interactions. For the target pathogen detection, we d eveloped surface -functionalization method to conjugate antibodies onto the MNP by physical non -covalent adsorption [160] [161] . We designed and optimized the process to label the target pathogen with MNP by antibody -antigen binding. Owning to signal penetration capability of the NMR, simple filtration method using a syringe filte r was found to be effective to remove excessive MNPs for both water and food samples. Then, the MNP concentration becomes proportional to the target pathogen concentration. Finally, the target pathogen concentration can be detected by the MNP -NMR biosenso r by measuring the NMR spin -spin relaxation time (T2) of water proton in the sample, which is inversely proportional to the MNP concentration. The MNP effect on NMR relaxation is described based on the general theory as follows. 103 5.2.9 Effects of MNP on NMR Relaxation Like electron s, the nucleus with unpaired protons or neutrons have an intrinsic quantum property of nuclear spin, which is characterized as quantized angular momentum, S. As circular current generates a magnetic field by Ampere™s law, a spinning nucle us also acts like a magnet of magnetic moment of defined in Equation 26, == Equation 26 where m is the magnetic quantum number, is the reduced Planck constant, and gyromagnetic ratio, a constant specific to the type of the nucleus. When placed in a magnetic field, nucleus would align along or against the field direction, creating a low energy state and a high energy state in Equation 27, == Equation 27 where m equals to 1/2 or -1/2 as quantum spin number for proton. The energy fibandgapfl between the high - and low - energy states equals to , , which is linearly proportional to , as illustrated in Figure 5.16. 104 Figure 5.16 The Zeeman effect: n uclei spin energies split when placed in an external magnetic field [168] . Similar to charged particles precess ing around a magnetic field, a charged nucleus, such as protons of water molecule s, will exhibit precession motion at a characteristic resonance frequency, called Larmor frequency, , which is determined by nucleus™s intrinsic quantum property and strength of the magnetic field where the sample was pl aced in (Equation 28). =/= Equation 28 The nucleus can be excited from low -energy state to high -energy state by absorbing energy from the oscillating magnetic field generated by the radio frequency (RF) signal at or close to Larmor frequency. 105 After the excitation, the nucleus continues to exchange energy with their surroundings, also called lattice, through thermal motion induced magnetic dipole -dipole interactions, and eventually b ring the population of spin energy state back to the thermal equilibrium. This process is called spin - lattice relaxation, or T1 relaxation, in which T1 denotes the duration. Similar with nucleus™ absorption of RF signal, the energy exchanges between the spin and lattice require the thermal motion/fluctuation occurring at a frequency at or close to the NMR resonance frequency. Thus, t he effectiveness of the process is determined by the number of the frequency matched population. For a simple case of unco upled spins of 1/2, the major source of T1 relaxation is dipolar coupling (DC), which is the interaction between adjacent nuclear spins, I and S, through space (Figure 5.17). The spin S induces a local field at the spin I, characterized by spin position function, (), as defined in Equation 29. ()=[3 ()1]()=() Equation 29 where () is the distance between spin I and S, and () is the angle between I-S axis and the external magnetic field, . 106 Figure 5.17 Dipolar interaction between nuclear spin s, I and S, through space . The spin S induces a local field at the spin I. Nuclei fluctuated by thermal motion experience different local field, causing Zeeman interaction fluctuating randomly with [169]. The physical effect of this random process can be described in the frequency domain by a spectral density function, (), which is the Fourier transform , defined in Equation 31, of an auto correlation function , () (Equation 30), for the randomly fluctuating local magnetic field, . As a result, () is the amount of motion at different frequencies , as derived in Equation 32. ()= (0) ()=(0)() Equation 30 Bloc B0 rIS I S S 107 ()=12() Equation 31 ()=2 J() Equation 32 where () is the function of the relative position interacting spins , the brackets represent a n averaging , and J() is the normalized spectral density function. The T1 of spin -lattice relaxation is the decay time of the net spin magnetization along the external magnetic field direction ( z-axis), Mz, that is reduced by the DC energy exchange. For two spin system, Mz can be modeled as Equation 33, =2(//) Equation 33 where / and / are the population at the two spin states, 1/2 and -1/2, respectively, and B is the Boltzmann factor, =. The kinetic of population variation of / and / are derived as Equation 34 and Equation 35, respectively. /=///+///) Equation 34 108 /=+//////) Equation 35 // and // are the transition probability between the two spin energy states defined as Equation 36and Equation 37 respectively , where P is the mean transition probability. //=(1/2) Equation 36 //=(1+/2) Equation 37 Thus, the kinetic of Mz can be obtained the last four equations as Equation 38, =2=2(1) Equation 38 Solving the above first order differential equation, the normalized () can be obtained as derived in Equation 39, ()=[(0)1]+1 Equation 39 Hence the relaxation rate, R1, or the reverse of T1 can be obtained from this exponential decay function as defined in Equation 40. =1=2 Equation 40 109 As aforementioned previously, () gives the amount of thermal motion at the resonance frequenc y, and hence determines the mean transition probability , as derived in Equation 41. =4()=2 J() Equation 41 Hence, the spin -lattice relaxation rate, R1, is then described by the spectral density at the spin resonance frequency, (), or the Fourier transform for the autocorrelation of spin position function, (), as derived in Equation 42. =2=2()=4() =4(0)() Equation 42 The mathematical form of the spectral density function J() is determined by the actual motional process. For isotropic diffusion, () is an exponential decay function, and is the motional correlation time (Equation 43). ()= / Equation 43 The derivation for spin -lattice relaxation time T1 can be expanded to describe the spin -spin relaxation time T2 of water proton influenced by s uperparamagnetic MNP, since both relaxations are of spin ½ and caused by fluctuating dipolar interaction. Their difference is that, the DC of the former is between the nuclear magnetic moment of p rotons, while the latter is between the proton ™s nuclear magnetic moment , and the MNP™s global magnetic moment , contributed by the exchange interaction, the Zeeman coupling, and the anisotropy energy . A superparamagnetic MNP has a 110 strong coupling between e lectronic spins , and its global magnetic moment is the total electronic magnetic moment of all the electrons in the crystal. Thus, the relaxation mechanism of water proton with MNP can be modeled based on the outer sphere theory, describing the effect of the interaction of the magnetic moment of a proton (I) with the magnetic moment of unpaired electrons (S), as derived in Equation 44 [170] [171] [172] . =1=18()()+()+3 4()[()+()]+3 4()()+() +2()() +32()() Equation 44 where is the Larmor frequency of the proton , () is the autocorrelation of the spin position function, (), as defined in Equation 45 and Equation 46, just like the case of proton dipole interaction as derived previously. ()=34(0)() Equation 45 111 ()=[3 ()1]() Equation 46 And = and = + are the raising and lowering operators of the Hamiltonian by quantum mechanics. () and () are the transverse spin -spin correlation functions and () is the longitudinal spin -spin correlation function. Superparamagnetic MNPs have very la rge anisotropy energy. They align with the external magnetic field, or the longitudinal axis. The contribution by longitudinal spin -spin correlation to relaxation is the square of the magnitude by the transverse correlation. Hence, the first three terms of the equation regarding transverse effect could be neglected with respect to the last two terms for longitudinal effect, and hence is simplified as Equation 47, =2()()+32()() Equation 47 The relaxation rate [172] , is obtained by solving the simplified equation above, which is in a general form similar to the T1 proton 1/2 spin case, but with more complexity describing the additional fluctuating dipole interaction with a MNP , as derived in Equation 48. 112 =3213500032(,,)+2(0,,)+322+2(0) Equation 48 where C is the MN P concentration . is the electron gyromagnetic ratio . is the proton gyromagnetic ratio . NA is the Avogadro constant. R is the effective radius of the MNP. And D is the water diffusion coefficient . is Freed™s spectral density function , which tak es into account for both the proton diffusion and the fluctuation of the magnetic moment around its mean value [173] . It is defined as shown in Equation 49. (,,)=1+//41+/+/9+//9 Equation 49 where =+/, =/ is the characteristic diffusion time, and is the Neel relaxation time of the MNP. is Ayant™s spectral density function accounting for the proton diffusion in a nonuniform magnetic field created by the mean electronic magnetic moment of the superparamagnetic MNP [174] . It is defined as Equation 50, as shown below . 113 ()=1+5/8+/81++/2+/6+4/81+/81+/648 Equation 50 The R2 equation above could be rewritten in terms of magnetic moment by substituting =/. Accordingly, becomes , which is project ed on the axis of the external magnetic field , . In this way, the R2 equation is better related to the actual physical mechanism. The first term by represents the relaxation induced by the fluctuating part of , or . The second term by describes the effect caused by the diffusion relat ed mechanism. Hence, R2 can be derived as Equation 51 [172] . =3213500032(,,)+2(0,,)+322+2(0) Equation 51 In a high external magnetic field, the MNPs™ magnetic vector s are fixed along the direction . Although the local fields of the MNPs are quite strong, their temporal variations in magnitudes are relatively small. Hence, the Curie relaxation dominates : the first term by can be neglected with respect to the second term by . Consider a si mple example for solid system, Equation 52, 114 == 11() Equation 52 where =tanh(), which is the expectation value of the magnetization , derived based on Curie™s law [175] . For our superparamagnetic MNP and NMR biosensor with the saturated magn etization of 38 emu/g and = 0.5 T, then =/=148.9 at T = 300 K. Finally, we obtained the ratio of =2.2×10. Thus, the effect by is considerably insignificant compared to that of . Finally , the spin -spin relaxation of w ater proton with superparamagnetic MNP in the liquid system are derived and simplified as shown in the Equation 53. =1/=32135000 322+2(0) Equation 53 The MNPs effectively induce nonuniformity of the NMR magnetic field in the sample, and hence enhance transverse magnetization dephasing and relaxation signal decay . The relaxation time , is inversely proportional to the concentration of the MNP, in linear relationship. With unique superparamagnetic MNPs synthesized for strong magnetization, antibody -MNP functionalization and conjugation methods to label the ta rget pathogen, we developed a new portable MNP -NMR biosensor system in this graduate research, which could be used to detect the target pathogen of different concentrations in water or food samples. 115 5.2.10 Detection of pNMR Relaxation Time The pNMR selects a pal m-sized permanent magnet of 0. 5 Tesla to facilitate portable application. However, c ompared to the bulky conventional super -conducting magnet, i t has trade -offs in magnetic field homogeneity and is challenging to implement a shim coil inside its limited internal space . Hence, the CPMG spin -echo technology is applied in the pNMR system to treat the less homogenous field . First, as with the free induction decay detection, the CPMG starts with an RF pulse applied to the sample B1 pulse) to rotate the magnetic dipole 90 degree from the field direction of the magnet (Figure 5.18). This results in the net magnetization being in the x -y plane or the transverse plane , leading to the maximum signal during the NMR relaxation . However, i n an inhomogeneous field, as determin ed by Equation 28, magnetic dipoles in different location s will have different La rmor frequenc ies , processing at different speeds, which causes the dipoles de-phasing and get out of step with each other in the transverse plane . This lead s to a rapid decay of the total magnetization vector after the 90 pulse is turned off , which caus es the NMR signal to decay too quickly , faster than the real spin -spin relaxation, T2. The CPMG method mitigate this problem by appl ying B1 pulse after waiting for the magnetization to decay away, at time [151] . B1 pulse reverse s the direction of the magnetic dipoles , causing them to change from de -phasing to refocus ing and eventually come back to step with each other . This will lead to a si gnal peak at time (a spin -echo) , reveal ing the actual peak as it were induced by T2 relaxation in a homogeneous field . 116 Figure 5.18 NMR spin -echo technique to treat the inhomogeneous field of a small permanent magnet using 90 degree and 180 degree RF pulse trains . (adapted and modified from [176] ) Based on Equation 23 B1 pulse can be achieved by applying the same oscillating B1 pulse. After the peak, the signal will start to decay rapidly, corresponding to magnetic dipoles™ de-phasing again. The Transverse Plane Mz Magnet Field Direction 117 so as shown in Figure 5.19. Figure 5.19 NMR spin -echo technique and CPMG pulse sequence to detect the spin -spin relaxation time, T2 in a less homogeneous magnetic field. As shown in Figure 5.1, a VHDL software was implemented at 300 MHz operation speed on a Xilinx Spartan -3 FPGA controller based on state machine to control the output switching of the power amplifier to generate t he CPMG sequence . The pulse frequency is determined by the magnet field strength (0.5 Tesla ) and proton™s gyromagnetic ratio, as 19.9 18 MHz. During the pulse -on time, the power amplifier is turned on to amplif y at full power for the 19.918 MHz pulse generated by the FPGA . During the pulse off time, the power amplifier was turned off to mini mize noise interreference to the receiver . The frequency, timing, width , duty cycle, and sequencing of pulse train were tested and verified using an oscilloscope . Through the free induction decay experiment on the pNMR system , the pulse widths of the 90 pulse and 180 pulse were tuned 180 RF Pulse 180 RF Pulse 1st spin-echo 2nd spin -echo 3rd spin -echo FID 180 RF Pulse 90 RF Pulse 2 2 2 180 RF Pulse RF Pulse NMR Signal Amplitude Time Time T2 Relaxation 90 spin 118 vely and optimized to be 1 .2 msec. The source code of the FPGA pulse control software is provided in Appendix A. 5.2.11 Test Pathogen and Antibodies Escherichia coli O157:H7 was obtained from the collection of the Nano -Biosensors Laboratory at Michigan State University. E. coli O157:H7 test strains were inoculated using a sterile loop into 10 mL of Tryptic soy nutrient broth from Difco Laboratories (Detroit, MI) and incubated for 24 h at 37 °C to make a stock culture. The stock culture was then serially diluted in 0.1% peptone water in logarithmic scale to obtain different concentrations. All the exp eriments were performed in a certified Biological Safety Level II laboratory. The antibody used for NMR biosensor was purified mouse anti -E. coli O157:H7 monoclonal antibody (Meridian Life Science, Inc. Saco, ME). 5.2.12 Magnetic Pathogen Separation Before pathogen conjugation, the antibody -MNP conjugates were filtered syringe filter (Millipore, MA, USA) to remove large particle s. Then, 50 µL of antibody - MNP conjugates and 50 µL of pathogen sample dilution were mixed in 400 µL of 0.01M PBS. For negative control (blank sample), 50 µl of 0.1% (w/v) peptone water was used instead of pathogen sample dilution. The solution was incubate d at 25 °C at 60 rpm for 30 minutes for MNP -pathogen conjugation [82] [177] . To enhance NMR sensitivity, the incubated solution was filtered using a syringe filter (Millipore, MA, USA) to remove impurities and unbound MNPs. The syringe filter™s pore size was determined to be as large as possible (0.45 µm) in order to ensure all the unbound particles c an flow through but keep blocking all the target bacteria. After a wash process using 0.01M PBS for 3 times , the syringe filter was backflushed using 5 mL of 0.01M PBS to release 119 the MNP -pathogen conjugates for further test using NMR. During separation, wa sh, and back -flush process, strong -field magnet was used to attract the MNPs in solution in order to facilitate the separation of unbound NMPs, hold the MNP -pathogen conjugates during washing, and help elute the MNP labeled pathogen , respectively . The filt er based magnetic separation process is illustrated in Figure 5.20. Figure 5.20 Working principle of the NMR based biosensor for pathogen detection 5.2.13 Sensor Architecture and Detection Principle After the filter based magnetic separation process, the interference of unbound MNPs was effectively reduced, and MNPs in the eluted solution were proportional to the path ogen concentration. As paramagnetic material, the MNPs induce d spatial and temporal disturbance in 120 the homogeneity and strength of the local magnetic field ( Figure 5.21). Due to the high surface area to volume ratio, this disturbance introduce d precession frequency variations in millions of protons of the surrounding water molecules, which accelerate d the decay of the spin system™s phase coherence. It was earlier shown that the MNP™s concentration has a linear relationship to the water proton™s spin -spin relaxation time, T2 [178] . Therefore, the concentration of target pathogen in test solution c ould be measured from T2 signal using the portable NMR biosensor. Figure 5.21 Magnetic nanoparticles as biomarker to detect the target pathogen by the NMR measurement . 5.2.14 Detection and Data Analysis The NMR biosensor signal was measured using a digital oscilloscope: Model Agilent DSO1024A (Agilent Technologies , Santa Clara , CA) the connected to the NMR™s signal output using BNC cable. For the biosensor test, a volume of 1 80 µL of the test solution by immunomagnetic NMR Relaxation versus Time 121 separation was applied to the biosensor . The NMR spin -echo relaxation signal was recorded by the oscilloscope, which is controlled by FPGA synchronization signal . The whole process of NMR relaxation was less than one minute . For data analysis, a minimum of three replications were performed for each experiment . All biosensors were c alibrated using a control sample which consisted of immunomagnetic -separated solution prepared using the same test solution but without the pathogen. Standard deviations and mean values for the data of each experiment were calculated using Excel . 5.3 Results and Discussion 5.3.1 Functionalization of Magnetic Nanoparticles with Antibod y The magnetic nanoparticle s used in pNMR biosensor are synthesized from amine functionalized Fe2O3 mag netic nanoparticle. The particle synthesis w as described i n Chapter 3. The MNPs were evaluated using transmission electron microscopy (TEM) and electron diffraction measurement . As indicated in the TEM image in Figure 5.22, the MNPs have spherical shape with a uniform average size of 80 nm and a total distribution between 50 to 100 nm [167] . The MNPs™ crystalline nature is confirmed from the electron diffraction rings as demonstrated in the inset of Figure 5.22. As confirmed earlier by M -H hysteresis measurement [167] , the MNP (1:0.6)™s saturation magnetization, M S, was found to be 38 emu/g at room temperature. The MNPs™ measured coercivity (HC) was measured to be the same as the unfunctionalized Fe2O3 MNPs , 180 Oe , indicating that its anisotropy magnetic energy was not affected by the polyaniline nano -shell functio nalization . 122 Figure 5.22 TEM image and electron diffraction image (inset) of the MNPs . 5.3.2 MNP Antibody Functionalization Pathogen -specific antibodies were functionalized onto MNP s by physical non-covalent adsorption. The Ab -MNP conjugation is mainly formed by electrostatic interaction between the negatively charged Fc portion of the antibodies and the positively charged polyaniline surface, along with other factors, including hydrophobic effect , electrostatic interaction, hydrogen bonding , and van der Waals interaction [167] [179] . The successful functionalization was confirmed by spectrophotometric studies earlier by our research team [167] . From measurement using UV-VIS scanning spectrophotometer, pure antibody solution had a characteristic wavelength peak of protein molecules at 280 nm. After the functionalization process with three times of magnetic separation, the supernatant solution had no peak at 280 nm, indicating that the antibodies were conjugated onto the MNPs effectively . 123 5.3.3 NMR R2 Relaxation Time Using the designed pNMR system, the target pathogen can be quantitively detected through measuring NMR spin -spin relaxation time of water in the sample, which is inversely proportional to the amount of captured MNPs, as described in Chapter 5.2.8 . To detect the concentration of the target pathogen , the NMR biosensor measures each test sample for the spin -spin relaxation time , T2 of water proton. First, t he pNMR transmits high -power RF excitation signal through the coil antenna to the test sample, which align the magnetic moment of water proton inside . During the relaxation process back to thermal equilibrium , energy was emitted as RF echo signals and detected by the coil antenna. The echo series have exponential decay characteristics , which can be modeled in Equation 54. ()= Equation 54 where M is the nuclear spin magnetization vector as a function of time, t, M0 is the initial nuclear spin magnetization vector, and T2 is the spin -spin relaxation time constant. 124 Figure 5.23 NMR biosensor relaxation signal of detection: (A) control (blank) sample , and (B) bacteria l sample The relaxation time T2 is calculated by curve fitting of the signal envelope of the NMR echo series . For example, two NMR relaxation signals for pathogen sample and control sample are presented in Figure 5.23. Figure 5.23 (B) is the NMR signal of a sample spiked with E. coli O157:H7 while Figure 5.23 (A) is the NMR signal of a sample with no bacteria (blank) . In this particular test , the spiked sample had a plate -count ed bacterial concentration of 226 CFU/mL. The curve fitting results are shown as dashed lines in Figure 5.23 with the corresponding equations are shown above. The envelope curve fitting of the bacteria sample in Figure 5.23 (B) resulted mathematically in the Equation 55. Based on Equation 54, the time constant T2 is 1/17.79, which equals to 0.0556 s or 55.6 ms. y = 1.994 e -17.97x Equation 55 125 For the control (blank) sample (0 CFU/mL ), the envelope curve fitting of Figure 5.23A resulted in the Equation 56. y = 2.099 e-9.734x Equation 56 The control sample consists of the same composition solution with the same amount of MNPs but without bacteria . It has a T2 relaxation time of is 1/9.734, which equal s to 0.1027 s or 102.7 ms (Figure 5.23B). The T2 time of the bacteria sample is 54.1% (55.6/102.7) shorter than the control sample. These data demonstrate that the NMR signal decays faster in contaminated samples than in samples with no bacteria due to the formation of magnetic clusters around the bacterial cell walls and correspondingly reducing the bulk spin -spin relaxation time of the nearby water molecules. The magnetic filtration process remov es the excessive unconjugated MNPs . This results in the captured MNPs being proportional to the target bacteria l concentration . This improves the NMR sensitivity when detecting samples with low bacteria l concentration as its composi tion is close to the blank sample . Whole milk and d rinking water and were used in the experiment to represent food samples. Milk and w ater were artificially inoculated with E. coli O157:H7 of concentration ranging from 10 1 to 107 CFU/mL. Magnetic f iltration and NMR detection followed the same procedure as described above. The pNMR biosensor results are measured being delta T2 of whole milk and drinking water samples are shown in Figure 5.24 and Figure 5.25. The curve fitting of signal envelope was processed for each experiment and T2 was calcu lated as described above. The sample average and variance of the delta T2 values , T2 Control T2 Sample , were plotted for the bacteria contaminated and control samples. At least three replicates were performed for each test sample. Figure 5.24 and 126 Figure 5.25 indicate that the relaxation times of all the bacteria contaminated samples are shorter than that of the control samples . For drinking water and milk samples, the delta T2 (Figure 5.24 and Figure 5.25) increases linearly with bacterial concentration from 10 1 CFU/mL up to 10 4 CFU/mL. The a ssociated P -values for test results of water and milk sample s were calculated and summarized in Table 5.5 and Table 5.6, respectively. The increase in T2 difference between the control and contaminated samples supports the formation of MNPs clusters conjugated with the target bacteria, resulting in a change in their nearby magnetic field , and affecting the nuclear spin of the proton atoms of the surrounding water molecules. The pNMR relaxation difference does not further increase when the bacterial concentration is 10 5 CFU/mL or higher. This effect could be attributed to several factors. First, in the case of high bacterial concentrations, the outer bacteria could form one or several layers of blocking shell s to the inner MNPs and hence reduce their impacts on the NMR relaxation due to the ir large size differences . MNPs are of around 80 nm diameter while E. coli O157:H7 is approximately in length. In addition , the biochemical reaction of MNP -bacteria conjugation could be decreased due to the probabilistic interactions between the bacteria and antibodies, antibody orientations on MNP surface, and stability of the bacteria -MNPs conjugates complex. Further, there could be insufficient amount of Ab-MNP s to capture all the bacteria in the sample, similar to a saturat ion effect. In fact, data show that the pNMR signals for bacterial concentration of 105 to 107 CFU/mL do not have much difference compare d to the pNMR signal of 10 4 CFU/mL. 127 Figure 5.24 The MNP -based pNMR biosensor™s measurement of relaxation time change, delta T2, for drinking water , which were contaminated by E. coli O157:H7. Table 5.5 P-value of NMR biosensor test results for E. coli O157:H7 in water sam ples Sample Pair (CFU/mL) P-value (n = 3) 101 vs 10 2 0.012 102 vs 10 3 0.013 103 vs 10 4 0.012 104 vs. 10 5 0.156 105 vs. 10 6 0.312 106 vs. 10 7 0.377 The pNMR biosensor result s for the whole milk samples are comparable to that of the drinking water samples , as demonstrated in Figure 5.25. The magnetic filtration assay has basically cleaned the sample matrix and removed excessive unbound MNPs . In general , the lowest detection limit 128 for milk and drinking water are down to the order of 10 1 CFU/mL or specifically, 92 CFU/mL and 76 CFU/mL for milk and water , respectively (p < 0.05, n = 3) . Bacteria can be detected with concentration from 10 1 CFU/mL to 10 7 CFU/mL. Figure 5.25 The MNP -based pNMR biosensor™s measurement of relaxation time change, delta T2, for whole milk samples , which were contaminated by E. coli O157:H7 . Table 5.6 P-value of NMR biosensor test results for E. coli O157:H7 in milk samples Sample Pair (CFU/mL) P-value (n = 3) 101 vs 10 2 0.010 102 vs 10 3 0.024 103 vs 10 4 0.031 104 vs. 10 5 0.131 105 vs. 10 6 0.185 106 vs. 10 7 0.698 129 The pNMR biosensor™s wide detection range is highly beneficial to food safety since certain organisms have varied infectious doses. For example, studies have shown that the infectious dose for some Shigella spp. is less than 10 organisms while that of toxigenic V. cholera is 10 4 organisms , which has one thousand times of difference [180] . The infective dose for another toxigenic pathogen, Salmonella is 10 3 organisms [181] . Results from the pNMR biosensor indicates that its detection sensitivity is better than those in the existing literature. A chip -NMR biosensor designed by Lee et al. has a detection sensitivi ty of 10 3 CFU/mL for Staphylococcus aureus [140] . This level of detection limit is consistent with our results when magnetic filtration is not applied , which is 103 CFU/mL for E. coli O157:H7 . 5.4 Conclusions This chapter describes a novel integrated design of an NMR biosensor, which make use of antibody -functionalized magnetic nanoparticle and filter based magnetic separation. The detection of the biosensor system s is fast, which includes a magnetic filtration assay of 20 min followed by a signal detection of 1 min. The average sensing limit for water and milk is 84 CFU/mL, lower than other NMR biosensor s reported in the literature. The linear range of the NMR biosensor is from 10 1 to 10 4 CFU/mL while the detection range spans from 10 1 CFU/mL to 10 7 CFU/mL. The detection application can be extended to other microbial or viral organisms by appropriate adaption for their corresponding antibodies. Hence , besides food safety application, the NMR biosensor described in this research has potential to be applied as rapid detection device s in food safety, biodefense, and clinical diagnostics. 130 This chapter is adapted from our recently published work in the Journal of Biological Engineering : Yilun Luo, and Evangelyn C. Alocilja. Portable Nuclear Magnetic Resonance Biosensor and Assay for a Highly Sensitive and Rapid Detection of Foodborne Bacter ia in Complex Matrices . Journal of Biological Engineering . 2017. 11(14). DOI: 10.1186/s13036 -017 -0053-8 131 CHAPTER 6 CONCLUSION AND FUTURE WORK 6.1 Conclusions In this dissertation , two biosensors based on nanofiber and nanoparticle NMR, respectively, were successfully developed to detect E coli O157:H7 bacteria in culture and in food samples. The work described in this dissertation also incorporated electrospinning technology and nanofiber surface treatment as well as NMR probe and transceiver design to improve their system performance for biosensing applications. The first research work in the dissertation focused on an electrospun nanofib er biosensor. A new type of electrospun n anofibrous membrane (ENM) was successfully synthesized for capillary flow assay (LFA) application. Conductive MNP s were synthesized and functionalized with antibody to be the biomarker and applied in magnetic separation for sample filtration. A novel LFA biosensor was developed using the ENM and the MNP. The ENM was optimized for capillary action and pathogen binding by improved fiber alignment, plasma treatment , and antibody conjugation. Owing to the unique nanostructure and higher surface area of ENM and MNP , the biosensor was capable of detecting E. coli O157:H7 from a 10 1 CFU/mL sample with linear response from 101 to 10 4 CFU/mL . The versatility of this biosensor was also evaluated. It was capable of detecting BVDV from a 103 CCID/mL sample, equivalent to 1/1000 of vir al concentration in infected bovine blood serum . The detection process was fast, detection time of 15 min from lateral flow process to data acquisition. Its application can be easily extended to detect other microbial or viral organisms 132 by appropriately changing the antibodies . This low -cost and simple measurement biosensor makes it possible for rapid field testing in food supply and healthcare . The second research work reported in this dissertation is a portable NMR biosensor. MNP s were synthesized and functionalized with antibodies for NMR biosensing. The research combine d an NMR system design for high signal -to-noise ratio and a unique filtrati on assay to improve sensitivity. The versatility of the biosensor was evaluated on both water and dairy food samples. The sample testing was rapid, including a magnetic filtration assay of 20 -30 min followed by a detection time of 1 min. The averaged detec tion limit on water and milk samples was 84 CFU/mL, lower than the other NMR biosensors reported in the existing literature. The biosensor was highly portable and sensitive in detection with linear response from 10 1 to 10 4 CFU/mL while total range from 10 1 to 10 7 CFU/mL. The biosensing application can be easily extended to detect other microbial or viral pathogens by adaption for the corresponding antibodies. Thus, in addition to food safety application, the NMR biosensor has demonstrated promising potential to be applied for rapid detection in healthcare diagnostics and biodefense. In summary, the electrospun - and pNMR - biosensors developed in this research have demonstrated sensitive detection performance and rapid response compared to conventional E. coli detection methods. Both of them can be used as a low cost and portable diagnosis system for on -field food and water testing, and be extended to detect ot her bacterial and viral pathogens. 133 6.2 Recommendations for Future Work Future work is recommended towards a number of research activities. For the electrospun biosensor , further improvement in the electrospinning process may be possible to increase the consistency in fiber alignment. Software algorithm to automatically process the impedance data is beneficial to further reduce detection time. Fellows et al. developed an LFA to detect glycoprotein CD4 [41] . Zheng et al. developed a n immunosensor to detect -Trophin protein [43] . Shi et al. develo ped a portable LFA biosensor to detect neomycin (NEO) and quinolones antibiotics (QNS) [44] . It is possible to implement the electrospun biosensor to detect other analytes, such as protein or IgG . Since t he EFM has demonstrated a promising potential in biosensing , it is possible to explore the feasibility of applying for another biosensor platform. For the pNMR biosensor, the magnetic field strength of the low -cost and small magnet shifts due to temperature varia tion , which can cause variance in the NMR result during on -field testin g. Possible temperature regulator can be applied to keep the system temperature suitable. It is also possible to implement a dynamic NMR frequency control to compensate the temperature effect on the NMR signal [182] . To further improve portability, all the NMR components can be installed into a single instrument enclosure. Lu et al. utilized m icrofluidic channels for sample transport, detection, and removal in NMR detection [59] . It is possible to integrate such device in the pNMR system to automate the detection pro cess. Gossuin et al. found that large magnetic field is beneficial to NMR detection [58] . Janis et al. utilized a 1.5 Tesla Halbach magnet for proton prepolarization [60] . Other small, high field strength, and low -cost magnet s can be evaluated for this biosensor, such as the Halbach array magnet, to exploit the possibility to further increase detection limit, reduce cost , and system size . 134 APPENDI X 135 APPENDIX This section contains the software source code in VHDL programing language on the FPGA control ler for the CPMG pulse control. ------------------------------------------------------------------------- -- nmr _biosensor w_cpm g_main.vhd ------------------------------------------------------------------------- -- Author: Yilun Luo -- Biosensors Lab, Michigan State University ------------------------------------------------------------------------- -- Description: This file tests the included UART component by -- sending data in serial form through the UART to -- change it to parallel form, and then sending the -- resultant data back through the UART to determine if -- the signal is corrupted or not. When the serial -- information is converted into parallel information, -- the data byte is displayed on the 8 LEDs on the -- system board. -- ------------------------------------------------------------------------- -- Revision History: -- 03/30/11 (LYL) Created -- 05/20/11 (LYL) Added a finite s tate machin e (FSM) for CPM G pulse gen eration ------------------------- ------------------------------------------------ library IEEE; use IEEE.STD_LOGIC_1164.ALL; --use IEEE.STD_LOGIC_ARITH.ALL; 136 use IEEE.STD_LOGIC_UNSIGNED.ALL; --use IEEE.numeric_bit.all; ------ YL use IEEE.numeric_std.all; -------------------------------- ----------------------------------------- -- Title: Main entity -- -- Inputs: 3 : RXD, CLK, RST -- -- Outputs: 1 : TXD, LEDS -- -- Description: This describes the main entity that tests the included -- UART component. The LEDS signals are used to -- display the data byte on the LEDs, so it is set equal to -- the dbOutSig. Technically, the dbOutSig is the scan code -- backwards, which explains why the LEDs are mapped -- backwards to the dbOutSig . ------------------------------------------------------------------------- entity DataCntrl is Port ( TXD : out std_logic := '1'; RXD : in std_logic := '1'; CLK : in std_logic; LEDS : out std_logic_vector(7 downto 0) := "11111111"; RST : in std_logic := '0'; --------------------------- YL------------------------- EnLED : out bit_vector(3 downto 0); EnDigit : out bit_vector(6 downto 0); DISPLAY : in std_logic; PulsOut : out std_logic := '0'; 137 BtnPuls : in std_logic := '0'; PULSRESET : in std_logic := '0'; SwSel : in std_logic_vector(3 downto 0); DeblankOut : out std_logic := '0'); --------------------------- YL------------------------- end DataCntrl; architecture Behavioral of DataCntrl is ------------------------------------------------------------------------- -- Local Component, Type, and Signal declarations. ------------------------------------------------------------------------- component RS232RefComp Port ( TXD : out std _logic := '1'; RXD : in std_logic; CLK : in std_logic; DBIN : in std_logic_vector (7 downto 0); DBOUT : out std_logic_vector (7 downto 0); RDA : inout std_logic; TBE : inout std_logic := '1'; RD : in std_logic; WR : in std_logic; PE : out std_logic; FE : out std_logic; OE : out std_logic; RST : in std_logic := '0' ); end component; 138 ------------------------------------ YL----------------------------------- - component BCDHEXDisplay is Port ( BINBCD: in std_logic; --0 for binary 1 for BCD BinNum: in bit_vector( 3 downto 0); DecNum: in integer; LEDIndex : in integer; LED: out bit_vector(3 downto 0); Digit: out bit_vector( 6 downto 0) ); end component; component LEDDisplayTiming is Port ( LEDCLK: in std_logic; EN: in std_logic; DATALED: in bit_vector(15 downto 0); BINBCD: out std_logic; LEDINDX: out integer range 0 to 5; LEDNUM: out bit_vector(3 downto 0) ); end component; com ponent Clock_Divider port ( CIN : in STD_LOGIC; TIMECONST1: in integer; COUT: out STD_LOGIC ); 139 end component; -- Debounce circuit for Key COMPONENT debounce PORT ( clk, key : IN STD_LOGIC; pulse : OUT STD_LOGIC); END COMPONENT; COMPONENT P_GENERATOR PORT( CLK : IN std_logic; -- CLKMS : IN std_logic; RESET : IN std_logic; TRIG : IN std_logic; PULSE : OUT std_logic; Deblank : OUT std_logic; WIDTH_A : in integer range 0 to 9999; WIDTH_B : in integer range 0 to 999025; WIDTH_C : in integer range 0 to 9999; -- WIDTH_D : in integer range 0 to 63 WIDTH_D : in integer range 0 to 127 ); END COMPONENT; ------------------------------------ YL------------------------------------ ------------------------------------------------------------------------- --- -- -- Title: Type Declarations -- -- Description: There is one state machine used in this program, called 140 -- the mainState state machine. This state machine controls -- the flow of data around the UART; allowing for data to be -- changed from serial to parallel, and then back to serial. -- ------------------------------------------------------------------------- type mainState is ( stReceive, stSend); ------------------------------------------------------------------------- ----------------------------------- YL------------------------- -------- type HFBYTE is array (0 to 3) of std_logic_vector(3 downto 0) ; ----------------------------------- YL--------------------------------- ------------------------------------------------------------------------- -- -- Title: Local Signal Declaratio ns -- -- Description: The signals used by this entity are described below: -- -- -dbInSig : This signal is the parallel data input for the UART -- -dbOutSig : This signal is the parallel data output for the UART -- -rdaSig : This signal will g et the RDA signal from the UART -- -tbeSig : This signal will get the TBE signal from the UART -- -rdSig : This signal is the RD signal for the UART -- -wrSig : This signal is the WR signal for the UART -- -peSig : This signal will get the PE si gnal from the UART -- -feSig : This signal will get the FE signal from the UART -- -oeSig : This signal will get the OE signal from the UART -- -- The following signals are used by the main state machine for state control: 141 -- -- -stCur, stNext -- ------------------------------------------------------------------------- signal dbInSig : std_logic_vector(7 downto 0); signal dbOutSig: std_logic_vector(7 downto 0); signal rdaSig : std_logic; signal tbeSig : std_logic; signal rdSig : std_logic; signal wrSig : std_logic; signal peSig : std_logic; signal feSig : std_logic; signal oeSig : std_logic; signal stCur : mainState := stReceive; signal stNext : mainState; ---------------------------------- YL---------------------------- --Signals for LED display signal LEDCLK: std_logic; signal LEDINDX: integer range 0 to 4:= 0; signal LEDNUM: bit_vector (3 downto 0); signal BINBCD: std_logic:='0'; signal ENDISPLAY: std_logic:= '1'; signal DATALED: bit_vector (15 downto 0); signal HEX IN: integer range 0 to 16:=0; --signal CLKDIV: integer range 0 to 10000000:=0; signal startPuls: std_logic:='0'; signal DEBCLK: std_logic; -- signal PULSCLK: std_logic; --signal PulsCount: std_logic_vector (4 downto 0); 142 ---- signal PulsCount: integer range 0 to 100000000; signal index: integer range 0 to 3 :=0; signal PulsWidthA: HFBYTE; signal PulsWidthB: HFBYTE; signal PulsWidthC: HFBYTE; signal PulsTimeA: integer range 0 to 9999; signal PulsTimeB: integer range 0 to 999025; signal Pu lsTimeC: integer range 0 to 9999; -- signal PulsTimeD: integer range 0 to 63 := 5; --Interval number of the longer pulse. -- signal PulsTimeD: integer range 0 to 63 := 40; --051112 signal PulsTimeD: integer range 0 to 127 := 80; --080312 -------- -------------------------- YL---------------------------- ------------------------------------------------------------------------ -- Module Implementation ------------------------------------------------------------------------ begin ------------------- ----------------------------------------------------- -- --Title: LED definitions -- ------------------------------------------------------------------------ -- LEDS(7) <= dbOutSig(0); -- LEDS(6) <= dbOutSig(1); -- LEDS(5) <= dbOutSig(2); -- LEDS(4) <= dbOutSig(3); -- LEDS(3) <= dbOutSig(4); 143 -- LEDS(2) <= dbOutSig(5); -- LEDS(1) <= dbOutSig(6); -- LEDS(0) <= dbOutSig(7); ------------------------------------------------------------------------- UART: RS232RefComp port map ( TXD => TXD, RXD => RXD, CLK => CLK, DBIN => dbInSig, DBOUT => dbOutSig, RDA => rdaSig, TBE => tbeSig, RD => rdSig, WR => wrSig, PE => peSig, FE => feSig, OE => oeSig, RST => RST); --------------- ---------------------------------------------------------- -------------------------------------- YL------------------------------------ L2: BCDHEXDisplay port map (BINBCD,LEDNUM,HEXIN,LEDINDX,EnLED,EnDigit); D1: LEDDisplayTiming port map (LEDCLK,ENDISPLAY,DATALED,BINBCD,LEDINDX,LEDNUM); -- Set countdown DebPuls: Debounce port map (DEBCLK, BtnPuls, startPuls); G1: Clock_Divider port map (CLK,100,LEDCLK); --2.5kHz. 10000 -> 400us G2: Clock_Divider port map (CLK,10000,DEBCLK); 144 --1MHz --G3: Clock_Divider port map (CLK,25,PULSCLK); -- Inst_P_GENERATOR: P_GENERATOR PORT MAP( -- CLK => CLK, -- CLKMS => PULSCLK, -- RESET => PULSRESET, -- TRIG => StartPuls, -- PULSE => PulsOut, -- Deblank => DeblankOut -- ); Inst_P_GENERATOR: P_GENERAT OR PORT MAP(CLK,PULSRESET,StartPuls,PulsOut,DeblankOut,PulsTimeA,PulsTimeB,PulsTimeC,Pul sTimeD); ---- DeblankOut <= StartPuls; -------------------------------------- YL------------------------------------ --------------------------------------------------- ---------------------- -- -- Title: Main State Machine controller (UART) -- -- Description: This process takes care of the Main state machine -- movement. It causes the next state to be evaluated on -- each rising edge of CLK. If the RST signal is strobed, -- the state is changed to the default starting state, which -- is stReceive. -- ------------------------------------------------------- ------------------ process (CLK, RST) begin 145 if (CLK = '1' and CLK'Event) then if RST = '1' then stCur <= stReceive; else stCur <= stNext; end if; end if; end process; ------------------------------------------------------ ------------------- -- -- Title: Main State Machine (UART) -- -- Description: This process defines the next state logic for the Main -- state machine. The main state machine controls the data -- flow for this testing program in order to send and -- receive data. -- ------------------------------------------------------------------------- process (stCur, rdaSig, dboutsig) begin case stCur is ------------------------------------------------------------------------- -- -- Title: stReceive state -- -- Description: This state waits for the UART to receive data. While in -- this state, the rdSig and wrSig are held low to keep the -- UART from transmitting any data. Once the rdaSig is set -- high, data has been received, and is safe to transmit. At 146 -- this time, the stSend state is loaded, and the dbOutSig -- is copied to the dbInSig in order to transmit the newly -- acquired parallel information. -- ------------------------------------------------------------------------- when stReceive => rdSig <= '0'; wrSig <= '0'; if rdaSig = '1' then --DATALED(7 downto 0) <= to_bitvector(dbOutSig(7 downto 0)); ---YL--- dbInSig <= dbOutSig; stNext <= stSend; else stNext <= stReceive; end if; ------------------------------------------------------------------------- -- -- Title: stSend state -- -- Description: This state tells the UART to send the parallel -- information found in dbInSig. It does this by strobing -- both the rdSig and wrSig signals high. Once these -- signals have been strobed high, the stReceive state is -- loaded. -- ---------------- --------------------------------------------------------- when stSend => rdSig <= '1'; 147 wrSig <= '1'; stNext <= stReceive; end case; end process; -------------------------------- YL--------------------------------------- -- p1: proc ess (rdaSig) -- begin -- if (rdaSig'event and rdaSig = '1') then -- -- if (Sw(7) = '1') then -- -- if (dbOutSig(7 downto 0)/="00001101") then ---YL: 'enter' -- case Command(15 downto ) is -- when "001" => -- PulsWidthA(index) <= dbOutSig(3 downto 0); -- when "010" => -- PulsWidthB(index) <= dbOutSig(3 downto 0); -- when "100" => -- PulsWidthC(index) <= dbOutSig(3 downto 0); -- when others => -- end case; -- ind ex <= index + 1; -- -- LEDS(3 downto 0) <= std_logic_vector( to_unsigned( index, 4)); -- DATALED(11 downto 8) <= "0101"; 148 -- --DATALED(7 downto 0) <= bit_vector( to_unsigned( index, 8)); -- else -- LEDS(3 downto 0) <= std_logic_v ector( to_unsigned( index, 4)); -- --DATALED(11 downto 8) <= "0111"; -- -- DATALED(3 downto 0) <= to_bitvector(PulsWidth(0)); -- DATALED(7 downto 4) <= to_bitvector(PulsWidth(1)); -- DATALED(11 downto 8) <= to_bitvector(PulsWidth(2)); -- DATALED(15 downto 12) <= to_bitvector(PulsWidth(3)); -- -- index <= 0; -- end if; -- else -- -- end if; -- -- end process p1; -------------------------------- YL--------------------------------------- p1: process (rdaSig) begin if (rdaSig'event and rdaSig = '1') then if (SwSel(3) = '1') then -- if (dbOutSig(7 downto 0)/="00001101") then ---YL: 'enter' 149 case SwSel(2 downto 0) is when "001" => PulsWidthA(index) <= dbOutSig(3 downto 0); when "010" => PulsWidthB(index) <= dbOutSig(3 downto 0); when "100" => PulsWidthC(index) <= dbOutSig(3 downto 0); when others => PulsWidth A(index) <= PulsWidthA(index); PulsWidthB(index) <= PulsWidthB(index); PulsWidthC(index) <= PulsWidthC(index); end case; index <= index + 1; LEDS(3 downto 0) <= std_logic_vector( to_unsigned( index, 4)); LEDS(7 downto 4) <= "0101"; --DATALED(7 downto 0) <= bit_vector( to_unsigned( index, 8)); else case SwSel(2 downto 0) is when "001" => DATALED(3 downto 0) <= to_bitvector(PulsWidthA(3)); DATALED(7 downto 4) <= to_bitvector(PulsWidthA(2)); DATALED(11 downto 8) <= to_bitvector(PulsWidthA(1)); DATALED(15 downto 12) <= to_bitvector(PulsWidthA(0)); 150 --PulsTimeA <= 875 + to_integer( unsigned(PulsWidthA(3)))+to_integer( unsigned(PulsWidthA(2)))*10+to_i nteger( unsigned(PulsWidthA(1)))*100+to_integer( unsigned(PulsWidthA(0)))*1000; PulsTimeA <= 805 + to_integer( unsigned(PulsWidthA(3)))+to_integer( unsigned(PulsWidthA(2)))*10+to_integer( unsigned(PulsWidthA(1)))*100+to_integer( unsigned(PulsWidthA(0 )))*1000; when "010" => DATALED(3 downto 0) <= to_bitvector(PulsWidthB(3)); DATALED(7 downto 4) <= to_bitvector(PulsWidthB(2)); DATALED(11 downto 8) <= to_bitvector(PulsWidthB(1)); DATALED(15 downto 12) <= to_bitvector(PulsWid thB(0)); --PulsTimeB <= to_integer( unsigned(PulsWidthB(3)))+to_integer( unsigned(PulsWidthB(2)))*10+to_integer( unsigned(PulsWidthB(1)))*100+to_integer( unsigned(PulsWidthB(0)))*1000; --PulsTimeB <= - 875 + to_integer( unsigned(PulsWidthB(3)) )*100+to_integer( unsigned(PulsWidthB(2)))*1000+to_integer( unsigned(PulsWidthB(1)))*10000+to_integer( unsigned(PulsWidthB(0)))*100000; PulsTimeB <= - 805 + to_integer( unsigned(PulsWidthB(3)))*100+to_integer( unsigned(PulsWidthB(2)))*1000+to_intege r( unsigned(PulsWidthB(1)))*10000+to_integer( unsigned(PulsWidthB(0)))*100000; when "100" => DATALED(3 downto 0) <= to_bitvector(PulsWidthC(3)); DATALED(7 downto 4) <= to_bitvector(PulsWidthC(2)); DATALED(11 downto 8) <= to_bitvector (PulsWidthC(1)); 151 DATALED(15 downto 12) <= to_bitvector(PulsWidthC(0)); --PulsTimeC <= 875 + to_integer( unsigned(PulsWidthC(3)))+to_integer( unsigned(PulsWidthC(2)))*10+to_integer( unsigned(PulsWidthC(1)))*100+to_integer( unsigned(PulsWidthC(0) ))*1000; PulsTimeC <= 805 + to_integer( unsigned(PulsWidthC(3)))+to_integer( unsigned(PulsWidthC(2)))*10+to_integer( unsigned(PulsWidthC(1)))*100+to_integer( unsigned(PulsWidthC(0)))*1000; when others => PulsTimeA <= PulsTimeA; PulsT imeB <= PulsTimeB; PulsTimeC <= PulsTimeC; end case; index <= 0; LEDS(3 downto 0) <= std_logic_vector( to_unsigned( index, 4)); LEDS(7 downto 4) <= "0111"; end if; end if; end process p1; ----------------------------------- YL--------------------------------- -- PulsOut <= startPuls; -------- YL: Debug Probe ------- -- PulsOut <= rdaSig; ----------------------------------- YL--------------------------------- --------------------------------- --YL--------------------------------- 152 -- PULS: process(CLK, CLKMS, startPuls) -- -- begin -- if (startPuls'event and startPuls = '0') then -- genPuls <= '1'; -- end if; -- -- if (genPuls = '0') then -- PulsCountA <= 0; -- PulsCountB <= 0; -- PulsCountC <= 0; -- elsif (CLK = '1' and CLK'Event) then ---- If (PulsCount /= 25) then -- If (PulsCount /= (PulsTimeA+PulsTimeC)) then -- PulsCount <= PulsCount + 1; -- end if; -- --end if; -- elsif (CLKMS = '1' and CLKMS'Event) th en -- If (PulsCountB /= (PulsTimeB)) then -- PulsCountB <= PulsCountB + 1; -- end if; -- end if; -- ---- if (startPuls = '1' and PulsCount < 24 ) then -- if (startPuls = '1') then -- if PulsCount<(PulsTimeA -1) then -- PulsOut <= '1'; -- elsif PulsCount<(PulsTimeA+PulsTimeB -1) then -- PulsOut <= '0'; -- elsif PulsCount<(PulsTimeA+PulsTimeB+PulsTimeC -1) then -- PulsOut <= '1'; 153 -- else -- PulsOut <= '0'; -- end if; -- else -- PulsOut <= '0'; -- end if; -- -- end process; ------------------------------------- YL--------------------------------- end Behavioral; 154 BIBLIOGRAPHY 155 BIBLIOGRAPHY [1] Atnafie B, Paulos D, Abera M, Tefera G, Hailu D, Kasaye S, Amenu K. Occurrence of Escherichia coli O157:H7 in cattle feces and contamination of carcass and various contact surfaces in abattoir and butcher shops of Hawassa, Ethiopia. BMC Microbiol. 2017 Jan 25;17(1):24. [2] Erickson MC, Liao JY, Payton AS, Cook PW, Ortega YR. Survival and internalization of Salmonella and Escherichia coli O157:H7 sprayed onto different cabbage cultivars during cultivation in growth chambers. J. Sci. Food Agric. 2019 May;99(7):3530 -3537. [3] Pollari F, Christidis T, Pintar KDM, Nesbitt A, Farber J, Lavoie MC, G ill A, Kirsch P, Johnson RP. Evidence for the benefits of food chain interventions on E. coli 0157:H7/NM prevalence in retail ground beef and human disease incidence: A success story. Can J Public Health. 2017 Apr 20;108(1):e71 -e78. [4] Federal, R., Quality s tandards for foods with no identity standards; bottled water , F.a.D. Administration, Editor. 1993. p. 52042 -52050. [5] Yakub, G.P., et al., Evaluation of Colilert and Enterolert defined substrate methodology for wastewater applications. Water Environment Research, 2002. 74(2): p. 131 -5. [6] Federal, R., Guidelines establishing test procedures for the analysis of pollutants; analytical methods for the biological pollutants in wastewater and sewage sludge; final rule. 2007. 72(157): p. 14220 -14233. [7] Fode -Vaughan, K.A., et al., Direct PCR detection of Escherichia coli O157:H7. Letters in Applied Microbiology, 2003. 37(3): p. 239 -43. [8] Torres -Chavolla, E. and E.C. Alocilja, Nanoparticle based DNA biosensor for tuberculosis detection using thermop hilic helicase -dependent isothermal amplification. Biosensors & Bioelectronics, 2011. 26(11): p. 4614 -8. [9] Hill, H.D. and C.A. Mirkin, The bio -barcode assay for the detection of protein and nucleic acid targets using DTT -induced ligand exchange. Nature Prot ocols, 2006. 1(1): p. 324 -36. [10] Rompre, A., et al., Detection and enumeration of coliforms in drinking water: current methods and emerging approaches. Journal of Micobiological Methods, 2002. 49: p. 31 -54. [11] Yanbo Liu, Ming Hao, Zhijun Chen, Lingling Liu, Ya o Liu, Wenxiu Yang, Seeram Ramakrishna, "A review on recent advances in application of electrospun nanofiber materials as biosensors," Current Opinion in Biomedical Engineering , vol. 13, pp. 174 -189, 2020. 156 [12] Ye WW, Ding YT, Sun Y,et al.: A nanoporous alumi na mem -brane based impedance biosensor for histamine detection with magnetic nanoparticles separation and amplification[J]. Procedia Technol 2017, 27:116 Œ117. [13] Xie S, Tang Y, Tang D,et al. :Highly sensitive impedimetric biosensor for Hg(2+) detection based on manganese porphyrin -decorated DNA network for precipitation poly -merization [J]. Anal Chim Acta 2018, 1023:22 Œ28. [14] Chowdhury AD, Ganganboina AB, Park EY,et al.: Impedimetric biosensor for detection of cancer cells employing carbohy -drate targeting ability of Concanavalin A[J]. Biosens Bio -electron2018,122:95 Œ103. [15] WHO (2017). Guidelines for drinking -water quality, fourth edition. World Health Organization: ISBN: 978 -92-4-154995 -0, pp. 127, 240 - 241, 2017. [16] CDC (2020). "Reports of Select ed E. coli Outbreak Investigations". Retrieved on 6 October 2020, from https://www.cdc.gov/ecoli/outbreaks.html [17] FDA (2020), " Spokane Produce Voluntarily Recalls Northwest Cuisine Creations and Fresh & Local Sandwiches and Green Leaf Lettuce Filets." Publ ished on 11 March 2020, from https://www.fda.gov/safety/recalls -market -withdrawals -safety -alerts/spokane -produce -voluntarily -recalls -northwest -cuisine -creations -and -fresh -local -sandwiches -and -green [18] Carol Holtz, fiGlobal Health Care: Issues and Policies.fl J ones & Bartlett Learning, pp. 239, May 20, 2016. [19] FDA (2020). Bacteriological Analytical Manual (BAM), U.S. Food & Drug Administration: Chapter 4a: Diarrheagenic Escherichia coli; Appendix 1: Rapid Methods for Detecting Foodborne Pathogens, 2020. [20] Guidelin es for drinking -water quality: fourth edition incorporating the first addendum. Geneva: World Health Organization; 2017. Licence: CC BY -NC-SA 3.0 IGO. [21] US National Institutes of Health (NIH), US National Institute of Allergy and Infectious Diseases (NIAID) . Biodefense - NIAID Category A, B, and C Priority Pathogens website. Available: http://www.niaid.nih.gov/topics/biodefenserelated/biodefense/pages/cata.aspx. Accessed 2012 Oct 26. [22] CDC (2018). "Bioterrorism Agents/Diseases." Retrieved on 4 April 2018, fro m https://emergency.cdc.gov/agent/agentlist -category.asp [23] Sujakhu, K. and U.T. Shrestha, Microbial (Coliform and Vibrio cholerae) analysis of drinking water collected from outbreak region of Bhaktapur municipality. 2011. 157 [24] CDC (2010). Microbiological Indicator Testing in Developing Countries: A Fact Sheet for the Field Practitioner. ver. 1 December 2010 https://sanitationupdates.files.wordpress.com/2010/11/microbiology2020.pdf [25] Bartram, Jamie, Ballance, Richard, World Health Organizatio n & United Nations Environment Programme. ( 1996). Water quality monitoring: a practical guide to the design and implementation of freshwater quality studies and monitoring programs / edited by Jamie Bartram and Richard Ballance. London: E & FN Spon. pp. 229-241, https://apps.who.int/iris/handle/10665/41851 [26] Sauerbrey G. Verwendung von schwingquarzen zur wägung dünner schichten und zur mikrowägung. Z. Phys. 1959;155:206 Œ222. doi: 10.1007/BF01337937. [27] Muratsugu M, Ohta F, Miya Y, Hosokawa T, Kurosawa S, Kam o N, Ikeda H, Quartz crystal microbalance for the detection of microgram quantities of human serum albumin: relationship between the frequency change and the mass of protein adsorbed. Anal Chem. 1993 Oct 15; 65(20):2933 -7. [28] Deng T, Li JS, Wang H, Shen GL, Yu RQ, Piezoelectric immunoassay for complement C4 based on a Nafion -modified interface for antibody immobilization. J Immunol Methods. 2005 Apr; 299(1 -2):1 -8. [29] Funari R, Terracciano I, Della Ventura B, Ricci S, Cardi T, D'Agostino N, Velotta R, Label -Free Detection of Gliadin in Food by Quartz Crystal Microbalance -Based Immunosensor. J Agric Food Chem. 2017 Feb 15; 65(6):1281 -1289. [30] Miroslav Pohanka, "Overview of Piezoelectric Biosensors, Immunosensors and DNA Sensors and Their Applications," Materials2018 ,11, 448; doi:10.3390/ma11030448 [31] Il-Hoon Cho, Dong Hyung Kim, and Sangsoo Park, "Electrochemical biosensors: perspective on functional nanomaterials for on -site analysis," Biomaterials Research (2020) 24:6. [32] Zhu N, et al. Label -free and sequence -specific DNA detection down to a picomolar level with carbon nanotubes as support for probe DNA. Anal Chim Acta. 2009; 650: 44 Œ8. [33] Han L, et al. Enhanced conductivity of rGO/Ag NPs composites for electrochemical immunoassay of prostate -specific antigen. Biosens Bio electron. 2017; 87: 466 Œ72. [34] Cai X, et al. Ratiometric electrochemical immunoassay based on internal reference value for reproducible and sensitive detection of tumor marker. Biosens Bioelectron. 2016;81:173 Œ80. [35] Damborský, Pavel et al. fiOptical biosensors .fl Essays in biochemistry vol. 60,1 (2016): 91 - 100. doi:10.1042/EBC20150010 158 [36] Huang, J.; Virji, S.; Weiller, B. H.; Kaner, R. B., Polyaniline Nanofibers: Facile Synthesis and Chemical Sensors. Journal of the American Chemical Society, 2003, 125, (2), pp. 314-315. [37] Sánchez -Tirado, E.; Salvo, C.; Gonz ález -Cort és, A.; Y áñez-Sede ño, P.; Langa, F.; Pingarr ón, J. M. Electrochemical Immunosensor for Simultaneous Determination of Interleukin -1 Beta and Tumor Necrosis Factor Alpha in Serum and Saliva Using D ual [38] Green, A. A.; Hersam, M. C. Properties and Application of Double -Walled Carbon Nanotubes Sorted by Outer -Wall Electronic Type. ACS Nano 2011, [39] Moore, K. E.; Flavel, B. S.; Ellis, A. V.; Shapter, J. G. Comparison of Double -Walled with Single -Walled Carbon Nanotube Electrodes byElectrochemistry.Carbon2011,49, [40] Hou, L. , Cui, Y. , Xu, M. , Gao, Z. , Huang, J. , Tang, D. , "Graphene Oxide -Labeled Sandwich -Type Impedimetric Immunoassay with Sensitive Enhancement Based on Enzymatic 4 -Chloro -1-Naphthol Oxidation ," Biosens. Bioelectron. 2013, [41] Tamika Fellows, Lance Ho, Shane Flanagan, Ronen Fog el, Dupe Ojo , and Janice Limson , " Gold nanoparticle -streptavidin conjugates for rapid and efficient screening of aptamer function in lateral flow sensors using novel CD4 -binding aptamers identified through Crossover -SELEX," Analyst, 2020,145, 5180 -5193 [42] Qiu, W., Baryeh, K., Takalkar, S. et al. Carbon nanotube -based lateral flow immunoassay for ultrasensitive detection of proteins: application to the determination of IgG. Microchim Acta 186, 436 (2019). https://doi -org.proxy1.cl.msu.edu/10.1007/s00604 -019 -3508-4 [43] Zheng, L., Fang, C., Yan, J. et al. A New Tactic for Label - -Trophin via Electrochemiluminescent Signalling on an AuNPs Supported Immuno -Interface. Sci Rep 7, 11199 (2017). ht tps://doi.org/10.1038/s41598 -017-11750 -8 [44] Shi, Q., Huang, J., Sun, Y. et al. A SERS -based multiple immuno -nanoprobe for ultrasensitive detection of neomycin and quinolone antibiotics via a lateral flow assay. Microchim Acta 185, 84 (2018). https://doi -org. proxy1.cl.msu.edu/10.1007/s00604 -017 -2556-x [45] Gondhalekar, C., Biela, E., Rajwa, B. et al. Detection of E. coli labeled with metal -conjugated antibodies using lateral -flow assay and laser -induced breakdown spectroscopy. Anal Bioanal Chem 412, 1291Œ1301 (202 0). https://doi -org.proxy1.cl.msu.edu/10.1007/s00216 -019-02347 -3 [46] Wiley: NMR Spectroscopy: Basic Principles, Concepts, and Applications in Chemistry, 2nd Edition - Harald Günther. [Online]. Available: 159 http://www.wiley.com/WileyCDA/WileyTitle/productCd -047195201X.html. [Accessed 14 Oct 2016]. [47] Ng SM. Chapter 3 - Portable NMR -Based Sensors in Medical Diagnosis A2 - ur-Rahman, Atta. In Applications of NMR Spectroscopy, M. I. Choudhary, Ed. Bentham Science Publishers, 2015, pp. 121 Œ146. [48] Yu D, Garcia N, Xu S. Toward portable nuclear magnetic resonance devices using a tomic magnetometers. Concepts Magn Reson Part A. 2009;34A(2): 124Œ32. [49] Perez JM, Josephson L, O™Loughlin T, Högemann D, Weissleder R. Magnetic relaxation switches capable of sensing molecular interactions. Nat Biotechnol. 2002;20(8):816 Œ20. [50] Lee H, Sun E, Ham D, Weissleder R. Chip ŒNMR biosensor for detection and molecular analysis of cells. Nat Med. 2008;14(8):869 Œ74. [51] Rokitta M, Rommel E, Zimmermann U, Haase A. Portable nuclear magnetic resonance imaging system. Rev Sci Instrum. 2000;71(11):4257 Œ62. [52] Lee H , Yoon T -J, Figueiredo J -L, Swirski FK, Weissleder R. Rapid detection and profiling of cancer cells in fine -needle aspirates. Proc Natl Acad Sci. 2009;106(30):12459 Œ64. [53] Zalesskiy SS, Danieli E, Blümich B, Ananikov VP. Miniaturization of NMR Systems: Deskt op Spectrometers, Microcoil Spectroscopy, and fiNMR on a Chipfl for Chemistry, Biochemistry, and Industry. Chem Rev. 2014;114(11): 5641Œ94. [54] Patel P. An NMR Chip The Size of a Seed. IEEE Spectrum: Technology, Engineering, and Science News, 05 -Aug -2014. [Onli ne]. Available: http:// spectrum.ieee.org/tech -talk/biomedical/devices/an -nmr -device -the -size -of-aseed. [Accessed 30 Sep 2016]. [55] Sara Hash, M. Pilar Martinez -Viedma, Fred Fung, Jee Eun Han, Paul Yang, Charlene Wong, Loganathan Doraisamy, Suresh Menon, Dona ld Lightner, "Nuclear magnetic resonance biosensor for rapid detection of Vibrio parahaemolyticus," Biomedical Journal, Volume 42, Issue 3, 2019, Pages 187 -192, ISSN 2319 -4170, https://doi.org/10.1016/j.bj.2019.01.009. [56] Suresh M. Menon, David E. Newman, Terry J. Henderson, J. Manuel Perez, " Magnetic resonance system and method to detect and confirm analytes," U.S. patent, US9442110B2, Sept. 13, 2016 [57] Gupta, M, Singh, K, Lobiyal, DK, et al. A sensitive on -chip probe -based portable nuclear magnetic resonance for detecting low parasitaemia plasmodium falciparum in human blood. Med Devices Sens. 2020; 3:e10098. https://doi .org.proxy1.cl.msu.edu/10.1002/mds3.10098 160 [58] Gossuin, Y., Okusa Ndjolo, P., Vuong, Q.L. et al. NMR relaxation properties of t he -hematin. Sci Rep 7, 14557 (2017). https://doi.org/10.1038/s41598 -017-15238-3 [59] Lu, R., Lei, P., Yang, Q. et al. Development of a Microfluidic NMR Device for Rapid and Quantitative Detection of Tumor Markers. Appl Magn Reson 50, 357Œ370 (2019). https://doi.org/10.1007/s00723 -018-1071-5 [60] Janis Smits, Joshua T. Damron, Pauli Kehayias, Andrew F. McDowell, Nazanin Mosavian, Ilja Fescenko, Nathaniel Ristoff, Abdelghani Laraoui, Andrey Jarmola, Victor M. Acosta, "Two -dimensional nucle ar magnetic resonance spectroscopy with a microfluidic diamond quantum sensor," Science Advances , Vol. 5, no. 7, 26 Jul 2019: eaaw7895 [61] Dengwei Jing, Le Sun, Jingyu Jin, Madasamy Thangamuthu and Junwang Tang, " Magneto -optical transmission in magnetic nanoparticle suspensions for different optical applications: a review ," Journal of Physics D: Applied Physics , Vol.54, Num, 1, October 14 2020. [62] Kai Wu, Jinming Liu, Diqing Su, Renata Saha, and Jian -Ping Wang, "Magnetic Nanoparticle Relaxation Dy namics -Based Magnetic Particle Spectroscopy for Rapid and Wash-Free Molecular Sensing," CS Appl. Mater. Interfaces 2019 , 11, 26, 22979Œ22986, Publication Date :June 6, 2019 , https://doi -org.proxy2.cl.msu.edu/10.1021/acsami.9b05233 [63] Ulrike Martens, Una Janke, Sophie Möller, Delphine Talbot, Ali Abou -Hassan, and Mihaela Delcea , " Interaction of fibrinogen Œmagnetic nanoparticle bioconjugates with integrin reconstituted into artificial membranes," Nanoscale , Vol . 12, Iss . 38, 09/2020, 19918-19930. [64] Kimberl y Fabijanczuk, Kaylee Gaspar, Nikunj Desai, Jungeun Lee, Daniel A. Thomas, J. L. Beauchamp, and Jinshan Gao, " Resin and Magnetic Nanoparticle -Based Free Radical Probes for Glycan Capture, Isolation, and Structural Characterization," Analytical Chemistry 2019 91 (24), 15387 -15396, DOI: 10.1021/acs.analchem.9b01303 [65] Shutyi, A.M., Eliseeva, S.V. & Sementsov, D.I. Responses of an Isolated Anisotropic Magnetic Nanoparticle and Nanoparticle Lattice to a Magnetic Field Pulse. Appl Magn Reson 51, 409Œ429 (2020). https://doi -org.proxy2.cl.msu.edu/10.1007/s00723 -020-01193 -2 [66] Pourmodheji H, Ghafar -Zadeh E, Magierowski S. A Multidisciplinary Approach to High Throughput Nuclear Magnetic Resonance Spectroscopy. Sensors. 2016;16 no. 6. [67] Lei K -M, Mak P -I, Law M -K, Martins RP. A palm - a digital microfluidic device and a semiconductor transceiver for chemical/biological diagnosis. Analyst. 2015;140(15):5129 Œ37. 161 [68] S. Kralj, M. Rojnik, Matija, J. Kos, D. Makovec (2013). "Targeting EGFR -overexpressed A431 cells with EGF -labeled silica -coated magnetic nanoparticles". Journal of Nanoparticle Research 15 (5), 1666, 1 -11. [69] F. Gertz1 and A. Khitun1 (2016). "Biological cell manipulation by magnetic nanoparticles". AIP Advances 6, 025308, 1 -7. [70] Y.S. Jiang, H. Wang, S.B. Li and W.J. Wen (2014). "Applications of Micro/Nanoparticles in Microfluidic Sensors: A Review". Sensors, 14, 6952 -6964 [71] I. Rabias, et al. (2010). "Rapid magnetic heating treatment by highly charged maghemite nanoparticles on Wistar rats exocranial glioma tumors at microliter volume". Biomicrofluidics 4(2): 024111 1 -8. [72] H. Unterwegera, S. Lyera, R.P. Friedricha, C. Jankoa, M. Pöttlera, S. Dürrb, C. Alexioua (2015). "Magnetic nanoparticle -based drug delivery for cancer therapy" . Biochemical and Biophysical Research Communications, Volume 468, Issue 3, Pages 463 Œ470. [73] K.E. Scarberry, E.B. Dickerson, J.F. McDonald, Z.J. Zhang (2008). "Magnetic Nanoparticle -Peptide Conjugates for in Vitro and in Vivo Targeting and Extraction of Can cer Cells". Journal of the American Chemical Society 130 (31): 10258 Œ62. [74] R. E. Elizabeth, J.P. Eric and E.S. Raymond (2013). "Magnetic separation of colloidal nanoparticle mixtures using a material specific peptide". Chem. Commun., 2013, 49, 5471 - 5473. [75] Xiliang, L., et al., Application of nanoparticles in electrochemical sensors and biosensors. Electroanalysis, 2006. 18(4): p. 319 -326. [76] Chen Q., Lin J., Gan C., Wang Y., Wang D., Xiong Y., Lai W., Li Y., Wang M. (2015) "A sensitive impedance biosensor based on immunomagnetic separation and urease catalysis for rapid detection of Listeria monocytogenes using an immobilization -free interdigitated array microelectrode". Biosens Bioelectron. 74:504 -11. [77] Pal, S., E.C. Alocilja, and F.P. Downes, (2007) Nanowire l abeled direct -charge transfer biosensor for detecting Bacillus species. Biosensors & Bioelectronics. 22(9-10): p. 2329 -36. [78] A.-H. Lu, E. L. Salabas and F. Schüth (2007). "Magnetic Nanoparticles: Synthesis, Protection, Functionalization, and Application". A ngew. Chem. Int. Ed. 46 (8): 1222 Œ1244 [79] S S.Rana, J. Philip, B.Raj (2010). "Micelle based synthesis of Cobalt Ferrite nanoparticles and its characterization using Fourier Transform Infrared Transmission Spectrometry and Thermogravimetry". Materials Chemistry and Physics 124: 264 Œ269. 162 [80] E. K. Athanassiou, Evagelos K.; R. N. Grass; W. J. Stark (2010). "Chemical Aerosol Engineering as a Novel Tool for Material Science: From Oxides to Salt and Metal Nanoparticles". Aerosol. Sci. Tech. 44 (2): 16 1Œ72. [81] Sharma, R., S. Lamba, et al. (2005). "Composition dependent magnetic properties of iron oxide -polyaniline nanoclusters." Journal of Applied Physics 97(1): 014311 -6. [82] Pal, S. and E. C. Alocilja (2009). "Electrically active polyaniline coated magnetic (EAPM) nanoparticle as novel transducer in biosensor for detection of Bacillus anthracis spores in food samples." Biosensors and Bioelectronics 24(5): 1437 -44. [83] Maghemite (2020), Retrieved 30 October 2020, from https://en.wikipedia.org/wiki /Maghemite. [84] K.S. Ryu, S.H. Chang, S.K. Jeong, E.J. Oh, and C.H. Yo. (2000). "Magnetic Properties of Conducting Polyanilines Induced by Organic Acids and Solvents". 238 Bull. Korean Chem. Soc., Vol. 21, No. 2, 238 -240 [85] Baric k, K. C., M. Aslam, et al. (2009). "Nanoscale assembly of amine -functionalized colloidal iron oxide." Journal of Magnetism and Magnetic Materials 321(10): 1529 -1532. [86] Peter Dobson, Stephen King, Erik Gregersen, and Kara Rogers, "Nanoparticle, " Encyclopedia Britannica , May 14, 2019, https://www.britannica.com/science/nanoparticle, Access November 30, 2020 [87] Lin, Y.Y., Wang, J., Liu, G.D., Wu, H., Wai, C.M., Lin, Y.H., 2008. Biosensors and Bioelectronics. 23, 1659 -1665. [88] Zhang, G.P., Wang, X.N., Yang, J.F., Yang, Y.Y., Xing, G.X., Li, Q.M., Zhao, D., Chai, S.J., Guo, J.Q., 2006. Journal of Immunological Methods. 312, 27 -33. [89] Beck, O., Kraft, M., Moeller, M.R., Smith, B.L., Schneider, S., Wennig, R, 2000. Ann Clin Biochem. 37, 199 -204. [90] Wong, R., Tse, H. , 2008. Lateral Flow Immunoassay, first ed. Humana Press, New York, pp. 35-51. [91] Basu, M., Seggerson, S., Henshaw, J., Jiang, J., Cordona, R., Lefave, C., Boyle, P.J., Miller, A., Pugia, M., Basu, S., 2004. Glycoconjugate Journal. 21, 487 -496. [92] Mehdinia, A. , Kazemi, S.H., Bathaie, S. Z., Alizadeh, A., Shamsipur M., Mousavi, M.F., 2009. Journal of Pharmaceutical and Biomedical Analysis. 49, 587 -593. [93] Claussen, J.C., Franklin A.D., Haque A.U., Porterfield, M.D., Fisher, T.S., 2009. ACS Nano. 3, 37 Œ44. 163 [94] Wang, X .Y., Drew, C., Lee, S. H., Senecal, K.J., Kumar, J., Samuelson, L.A., 2002. Journal of Macromolecular Science. A39, 1251 -1258. [95] Huang, Z.M., Zhang, Y.Z., Kotaki, M., Ramakrishna, S., 2003. Composites Science and Technology. 63, 2223 -2253. [96] Kim, I.D., Roths child, A., Lee, B.H., Kim, D.Y., Jo, S.M., Tuller, H.L., 2006. Nano Lett. 6, pp. 2009Œ2013. [97] Li, M.Y., Guo, Y., Wei, Y., MacDiarmid, A.G., Lelkes, P.I., 2006. Biomaterials, 27, 2705 Œ 2715. [98] Zeng, J., Xu, X.Y., Chen, X.S., Liang, Q.Z., Bian, X.C. , Yang, L.X ., Jing, X.B., 2003. Journal of Controlled Release. 92, 227 -231. [99] Venugopal, J.R., Low, S., Choon, A.T., Kumar, A. B., Ramakrishna, S., 2008. Artificial Organs. 32, 388 Œ 397. [100] Nama, J.H., Joo, Y.H., Lee, J.H., Chang, J.H., Cho, J.H., Chun, M.P., Kim, B.I ., 2009. Journal of Magnetism and Magnetic Materials. 321, 1389 Œ1392. [101] Hou, S., Li, X.X., Li, X.Y., Feng, X.Z., 2009. Biofabrication. 1, 1 -7. [102] Wang, Z.G., Wan, L.S., Liu, Z.M., Huang, X.J., Xu, Z.K., 2009. Journal of Molecular Catalysis B: Enzymatic. 56, 189 Œ195. [103] Ren, G.L., Xu, X.H., Liu, Q., Cheng, J., Yuan, X.Y., Wu, L.L., and Wan, Y.Z., 2006. Reactive & Functional Polymers. 66, 1559 Œ1564. [104] Sawicka. K., Gouma, P., Simon, S., 2005. Sensors and Actuators B. 108, 585 Œ588. [105] Quadrat, O., Stejs kal, J., Kraochvil, P., Klason, C., McQueen, D., Kubat, J., Saha, P., 1998. Syn. Metal. 97, 37 Œ42. [106] Tahir, Z.M., Alocilja, E.C., 2003. Biosensors and Bioelectronics. 18, 813 -819. [107] Kim, J.H., Cho, J.H., Cha, G.S., 2000. Biosensors and Bioelectronics. 14, 90 7Œ915. [108] Brock, K. V., Chase, C. C. L., 2000. Veterinary Microbiology. 77, 209 -214. [109] Nartker S., L. T. Drzal, and M. Misra (2005). fiAligned Electrospun Nanofibers for Nanocomposite Applications,fl Proceedings of the American Society for Composites, pp. 77-89. [110] O. Quadrat, J. Stejskal, P. Kraochvil, C. Klason, D. McQueen, J. Kubat, and P. Saha (1998). fiElectrical properties of polyaniline suspensions,fl Syn. Metal, vol. 97, pp. 37 Œ42. 164 [111] S. Ramakrishna, K. Fujihara, W.E. Teo, T.C. Lim and Z.W. Ma, (2005) "An Int roduction to Electrospinning and Nanofibers", World Scientific, pp 142 -145. [112] Z. M. Tahir, E. C. Alocilja (2003). fiA conductometric biosensor for biosecurity,fl Biosensors and Bioelectronics, vol. 18, pp. 813 -819. [113] ICTVdB - The Universal Virus Database, version 4. Retrieved 11 December 2010, from http://www.ncbi.nlm.nih.gov/ICTVdb/ICTVdB/. [114] Ridpath, J. F. (2010). Bvdv: Detection, Risk Management and Control. Proceedings of Texas Veterinary Medical Association, p. 151 -158. [115] Cloutier BC, Cloutier AK, Alocilja EC. Optimization of electrically active magnetic nanoparticles as accurate and efficient microbial extraction tools. Biosensors (Basel) . 2015;5(1):69 -84. Published 2015 Feb 5. doi:10.3390/bios5010069. [116] J. H. Kim, J. H. Cho, and G. S. Cha (2000 ). fiConductimetric membrane strip immunosensor with polyaniline -bound gold colloids as signal generator,fl Biosensors & Bioelectronics, vol. 14, pp. 907 Œ915. [117] K. V. Brock, and C. C. L. Chase (2000). fiDevelopment of a fetal challenge method for the evaluatio n of bovine viral diarrhea virus vaccines,fl Veterinary Microbiology, vol. 77, iss. 1-2, pp. 209-2140. [118] J.C. Chiang, A.G. MacDiarmid (1986). fiPolyaniline: protonic acid doping of the emaraldine form to metallic regime,fl Synthetic Metal, vol. 13, pp. 193 -205. [119] M.P. Doyle, Food Safety: Bacterial Contamination, Editor(s): Benjamin Caballero, Encyclopedia of Human Nutrition (Third Edition), Academic Press, 2013, Pages 322 -330, ISBN 9780123848857, https://doi.org/10.1016/B978 -0-12-375083 -9.00124-0. [120] Shi L, Xu L, Xiao R, Zhou Z, Wang C, Wang S and Gu B (2020) Rapid, Quantitative, High -Sensitive Detection of Escherichia coli O157:H7 by Gold -Shell Silica -Core Nanospheres -Based Surface -Enhanced Raman Scattering Lateral Flow Immunoassay. Front. Microbiol. 11:596005. doi: 10.3389/fmicb.2020.596005. [121] Chunxia Lu, Xiaoxu Gao, Ya Chen, Jiangtao Ren & Changbin Liu (2020) Aptamer -Based Lateral Flow Test Strip for the Simultaneous Detection of Salmonella typhimurium , Escherichia coli O157:H7 and Staphylococcus aureus , Analytical Letters, 53:4, 646-659, DOI: 10.1080/00032719.2019.1663528 . [122] Zonghan Wang, Xiaolin Yao, Yongzhi Zhang, RongWang , Yanwei Jia , Jing Sun , Daohong Zhang , JianlongWang , (2020) Functional nanozyme mediated multi -readout and label -free lateral flow immunoassay for rapid detection of Escherichia coli O157:H7, Food Chemistry , Vol . 329, 1 November 2020, 127224 . 165 [123] A. Hatch, B. Weigl, D. Zebert, and P. Yager (1999). fiMicrofluidic approaches toward immunoassays,fl in Proc. SPIE Microfluidic Devices Syst. II, Santa Clara, CA. [124] G. Cavelier (1995). fiPossible role of surface electrochemic al electron -transfer and semiconductor charge transport processes in ion channel function,fl Bioelectron. Bioeng., vol. 40, pp. 197 Œ213. [125] Günther H., 1995, NMR Spectroscopy: Basic Principles, Concepts, and Applications in Chemistry, 2nd ed., Wiley, 15 -48. [126] G.W. Ewing, 1997, Analytical Instrumentation Handbook, 2nd ed., CRC, 811 -812. [127] Doscotch M.A., Evans J.F., Munson E.J., 1998, Journal of Chemical Education, 75, 8, 1008-1013. [128] Takeda K., 2007, Rev. Sci. Instrum., 78, 033103. [129] Epel B., Sundramoorthy S.V., Mailer c., Halpern H.J., 2008, Concepts Magn Reson Part B Magn Reson Eng, 33B, 3, 163 Œ176. [130] Takeda K., 2008, Journal of Magnetic Resonance, 192, 218 Œ229. [131] Hassibi A., Babakhani A., Hajimiri A., 2009, IEEE Journal of Solid -State Circuits, 44, 6, 1805-1813. [132] Yong Liu, Nan Sun, Hakho Lee, Ralph Weissleder, and Donhee Ham, " CMOS mini nuclear magnetic resonance system and its application for biomolecular sensing," Proceedings of IEEE International Solid -State Circuits Conference , pp. 140 -141 (Feb 2008) [133] Sun N. , Yoon T.J., Lee H., Andress W.F., Weissleder R., Ham D., 2011, IEEE Journal of Solid -State Circuits (JSSC), 46, 1, 342 -352. [134] Sun N., Liu Y., Lee H., Weissleder R., Ham D., 2009, IEEE Journal of Solid -State Circuits (JSSC), 44, 5, 1629 -1643. [135] ZYSW -2-50DR+, Reflective SPDT, Solid State Switch, https://www.minicircuits.com/WebStore/dashboard.html?model=ZYSW -2-50DR%2B, June 2020 [136] Basys 2 Spartan -3E FPGA Trainer Board, https://store.digilentinc.com/basys -2-spartan -3e-fpga -trainer -board -limited -time/, June 2020 [137] 50B High Frequency (HF) solid state power amplifier, Henry Radio Inc., https://www.henryradio.com/amplifiers.html, June 2020 166 [138] AU-1467 low noise amplifier, "Bipolar amplifiers," Narda Miteq Inc., retrieved in June 2020, https://nardamiteq.com/docs/C22E.PDF. [139] Sonoma -instrument 310 low noise amplifier, "Sonoma Instrument 310 Broadband Amplifier," retrieved in September 2020, http:/ /sonoma -instrument.com/pdf/sonoma -instrument -310.pdf [140] Lee H., Sun E., Ham D., Weissleder R., 2008, Nature Medicine, 14, 8, 869 -874. [141] CDC, 2009, E. coli Outbreak Investigations, http://www.cdc.gov/ecoli/2009/1124.html. [142] Dundas S., Todd W.T., Stewart A.I., Murdoch P.S., Chaudhuri A.K., Hutchinson S.J., 2001, Clin Infect Dis, 33, 7, 923 -931. [143] CDC, 2011, E. coli Outbreak Investigations, http://www.cdc.gov/ecoli/outbreaks.html. [144] Frenzen P.D., Drake A., Angulo F.J., 2005, J Food Prot, 68, 12, 2623 -2630. [145] Wells J .G., Davis B.R., Wachsmuth I.K., 1981, Journal of clinical microbiology , 18, 3, 512Œ520. [146] March S. B., Ratnam S., 1986, J Clin Microbiol, 23, 5, 869 -872. [147] Guan J., Levin R.E., 2002, Food Biotechnology, 16, 3, 155 Œ 166. [148] Fratamico P.M., Strobaugh T.P., Med ina M.B., Gehring A.G., 1998, Biotechnology Techniques, 12, 7, 571 -576. [149] Lin Y.H., Chen S.H., Chuang Y.C., Lu Y.C., Shen T.Y., Chang C.A., Lin C.S., 2008, Biosensors and Bioelectronics, 23, 1832 Œ1837. [150] Permanent magnets PM1055, Metrolab Inc. , Geneva, Switz erland ( 2011) http://www.metrolab.com/index.php?id=28. [151] Fukushima E., 1993, Experimental Pulse NMR: A Nuts and Bolts Approach. [152] Behzad Razavi, 2011, RF Microelectronics (2nd Edition), Prentice Hall Press, pp.74 [153] Selectivity and Q of a Circuit, http://hyperphysics.phy -astr.gsu.edu/hbase/electric /serres.html, 2020 [154] Louis N. Hand and Janet D. Finch, "Analytical Mechanics," Cambridge, England: Cambridge University Press, pp. 192. ISBN 978 -0-521-57572 -0, 1998. 167 [155] Fundamental Physical Constants, "Proto n gyromagnetic ratio in MHz/T", NIST, https://physics.nist.gov/cgi -bin/cuu/Value?gammapbar, 2020. [156] D. I Hoult and R. E Richards. The signal -to-noise ratio of the nuclear magnetic resonance experiment. Journal of Magnetic Resonance (1969), 24(1):71 Œ85, Octo ber 1976. 1.3.1, 1.3.1.1 [157] R.C. Hansen, "Fundamental limitations in antennas," Proc. of IEEE , 69(2), pp. 170 -182, 1981. [158] A.D. Yaghjian, andS.R. Best, "Impedance, bandwidth, and q of antennas," IEEE Trans. Antennas Propagat. , 53(4), pp. 1298 -1324, 2005 [159] Leon g, Kenneth, J. Mazierska, Mohan V. Jacob, and D. Ledenyov. "Comparing unloaded Q-factor of a high -Q dielectric resonator measured using the transmission mode and reflection mode methods involving S -parameter circle fitting." In 2002 IEEE MTT -S Internationa l Microwave Symposium Digest (Cat. No. 02CH37278) , vol. 3, pp. 1665 -1668. IEEE, 2002. [160] Luo and Alocilja, Portable nuclear magnetic resonance biosensor and assay for a highly sensitive and rapid detection of foodborne bacteria in complex matrices, Journal o f Biological Engineering (2017) 11:14 [161] EC Alocilja, Y Luo, Nuclear magnetic resonance apparatus, systems, and methods, US Patent 10,359,378, 2019 [162] Harald Günther, NMR Spectroscopy: Basic Principles, Concepts, and Applications in Chemistry, Wiley 2nd Edition, 14 Oct 2016 [163] Ng SM. Chapter 3 - Portable NMR -Based Sensors in Medical Diagnosis A2 - ur-Rahman, Atta. In Applications of NMR Spectroscopy, M. I. Choudhary, Ed. Bentham Science Publishers, 2015, pp. 121 Œ146. [164] Yu D, Garcia N, Xu S. Toward portable n uclear magnetic resonance devices using atomic magnetometers. Concepts Magn Reson Part A. 2009;34A(2): 124 Œ32. [165] Conventional nuclear magnetic resonance (NMR) system, https://en.wikipedia.org -/wiki/Nuclear_magnetic_resonance, Accessed Sept 2019 [166] Stefan H Bo ssmann, Hongwang Wang, Magnetic Nanomaterials: Applications in Catalysis and Life Sciences, Royal Society of Chemistry, Cambridge 2017 [167] Pal S ., Setterington E, Alocilja EC. Electrically active magnetic nanoparticles for concentrating and detecting Bacillus anthracis spores in a direct -charge transfer biosensor. IEEE Sensors J. 2008;8(6):647 Œ54. 168 [168] Nuclei spin energies split in an external magnetic field, https://en.wikipedia.org/wiki/Nuclear_magnetic_resonance/media/File:NMR_splitting.gif, , Accessed Oct 2019 [169] Dybowski, C. Zeeman Interaction in Nuclear Magnetic Resonance. In Encyclopedia of Analytical Chemistry (eds R.A. Meyers and C. Dybowski). doi:10.1002/9780470027318.a6119, 15 September 2006 [170] Gueron M (1975). Nuclear relaxation in macromolecules by paramagnetic ions: A novel mechanism. J Magn Reson 19, 58 Œ66. [171] L. Banci, I. Bertini, and C. Luchinat (1991). Nuclear and Electron Relaxation, VCH, Weinheim . [172] Roch, Muller, and Gillis, Theory of proton relaxation induced by superparamagnetic particles, J. Chem. Phys., Vol. 110, No. 11, pp. 5403 -5411, 15 March 1999 [173] Freed JH (1978). Dynamic effects of pair correlation functions on spin relaxation by translational diffusion in liquids . J Chem Phys 68, 4034 Œ4037. [174] Y. Ayant , E. Belorizky, J. Aluzon, Jean Gallice. Calcul des densites spectrales resultant d'un mouvement aleatoire de translation en relaxation par interaction dipo - laire magnetique dans les liquides. Journal de Physique, 1975, 36 (10), pp.991 -1004. [175] Kittel, Charles. Introduction to Solid State Physics, 8th Edition. Wiley. pp. 304, 2004 [176] Apostolos Kantzas, Jonathan Bryan, Saeed Taheri, " Fundamentals of Fluid Flow in Porous Media," retrieved in October 2020, https://perminc.com/resources/fundamentals -of-fluid -flow -in-porous -media/chapter -3-molecu lar -diffusion/diffusion -coefficient/measurement -techniques/nmr -method/principles -nmr -processing/spin -echo -cpmg -pulse -sequence/ [177] Luo, Y., Nartker, S., Miller, H., Hochhalter, D., Wiederoder, M., Wiederoder, S., Setterington, E., Drzal, L.T., Alocilja, E.C., 2010, Biosensors and Bioelectronics, 26, 4, 1612-1617 [178] Roch, A., Muller, R.N., and Gillis, P. (1999) Theory of Proton Relaxation Induced by Superparamagnetic Particles, J. Chem. Phys. 110, 5403Œ5411. [179] Sharma S, Byrne H, O™Kennedy RJ. Antibodies and Antibo dy-Derived Analytical Biosensors. Ed. Pedro Estrela. Essays Biochem. 2016;60(1):9 Œ18. [180] Kothary MH, Babu US. Infective Dose of Foodborne Pathogens in Volunteers: A Review. J Food Saf. 2001;21(1):49 Œ68. [181] Loynachan AT, Harris DL. Dose Determination for Acute Salmonella Infection in Pigs. Appl Environ Microbiol. 2005;71(5):2753 Œ5. 169 [182] E.C. Alocilja, Y . Luo, Nuclear magnetic resonance apparatus, systems, and methods, US Patent 10,359,378, 2019