/

University of Southern California Dissertations and Theses

/

Integration of practical high sensitivity whispering gallery mode resonator sensors

(USC Thesis Other)

Integration of practical high sensitivity whispering gallery mode resonator sensors

PDF

Transcript (if available)

Content INTEGRATION OF PRACTICAL HIGH SENSITIVITY WHISPERING GALLERY MODE RESONATOR SENSORS by Thanh M. Le A Dissertation Presented to the FACULTY OF THE USC GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (ELECTRICAL ENGINEERING) May 2012 Copyright 2012 Thanh M. Le ii Acknowledgements I want to express my deepest gratitude to all the people that helped made this dissertation a success. Without their significant support, this research would have not been possible. I would like to take this time to thank them. First and foremost I would like to thank my advisors, Prof. W.H. Steier and Dr. Lute Maleki, for their guidance in my studies. Their great knowledge, leadership, experience, and motivation has helped me become a better engineer and researcher. Their insight and ingenious advice has helped guide my research in the right direction. I have truly learned a lot during my thesis, which goes beyond science itself. Furthermore, I would like to thank all my professors from the Electrical Engineering and Electro-Physics Department at the University of Southern California, from whom I have received great knowledge. From these Departments, I would especially like to thank the members of my Dissertation Committee: Prof. Sawchuk and Prof. Armani. In addition, I would like to express gratitude to Prof. Hellwarth, Prof. Prata and Prof. O’Brien for their presence as my Qualifying Exam Committee. Also, I would like to thank all the former and current members of my research group at USC: Dr. Hidehisa Tazawa, Dr. Reem Song, Dr. Bipin Bhola, Dr. Greeshma Gupta, Dr. Andrew Yick, Dr. Yoo Seung Lee, Satsuki Takahashi and Hagi Mahalingam. Another group of people I would also like to thank are all of my co-workers in the Quantum and Science Technology Group at the Jet Propulsion Lab (JPL) for their advice and useful discussions. I also want to express my appreciation towards my previous members from JPL, Dr. Vladimir Ilchenko, Dr. Andrey Mastko, Dr. Anatoliy iii Savachenkov and Dr. Makan Mohageg, who are now at OEWaves Inc. in Pasadena, CA, for their help and support through training in the laboratory or helpful discussions. A special thank you goes out to Dr. Nan Yu, my group supervisor, for all of his support and encouragement during all my research projects for work or for my doctorate. Finally, I want to acknowledge and thank my family who has been very important to me throughout my life. Their dedication and encouragement has allowed me to continue my education to the doctorate level. Many thanks go to my sisters, Nhien and Mai, and their families, the Nyman family, and the Do family for believing and supporting me. Nevertheless, I am most grateful to my wonderful wife, Kim, and my amazing daughters, Melissa and Annabel for their continuous support throughout my doctoral research, encouragement when the days are hard, and simply bringing joy and inspiration to my life. My loving wife always has stood by me during the times of frustration. Her patience, understanding, and unconditional support have helped me get through some of most difficult challenges in the pursuit of my career. I am blessed by her love every day, and I want to thank her for all the sacrifices she has made for my daughters and I. Most importantly, I dedicate this doctoral degree to my parents, Trong Le and Tam Nguyen, who sacrificed their life of comfort to bring me to the country with the best education program, in order to give me the incredible opportunity to pursue a career in science. Their unconditional love, encouragement, unwavering trust, and constant support have influenced me into the person I am today. I am forever grateful to them. In addition, I especially want to thank and dedicate this work in memory of my brother-in-law, iv Leslie Nyman, M.D., who supported and encouraged me to embark on this journey from the day I stepped foot in this country. v Table of Contents Acknowledgements ii List of Tables vii List of Figures viii Abstract xiv Chapter 1: Introduction 1 1.1 Background and Motivation 1 1.2 Dissertation Outline 5 1.3 Chapter References 7 Chapter 2: Whispering Gallery Mode Resonators 8 2.1 Introduction 8 2.2 Geometric Optics Analysis of WGM Resonator 13 2.3 Analytic Approximation for The Modes in a WGM Disk Resonator 14 2.4 Properties of WGM Resonator 22 2.4.1 Quality Factor Q and System Loss 22 2.4.2 The Free Spectral Range 26 2.4.3 Finesse 28 2.5 Chapter References 29 Chapter 3: Optical Polymer Waveguide Coupled to High Q Whispering Gallery Mode Resonators 32 3.1 Introduction 32 3.2 Theoretical Approach of Polymer Waveguides Coupled to WGM Resonators 33 3.3 Design and Fabrication Optical Polymer Waveguides and High Q WGM Fused Silica Resonator Disks 36 3.4 Experiment and Results 44 3.5 The Integrated Devices Demonstration 49 3.6 Chapter References 52 Chapter 4: Differential TE and TM Mode Measurement: A Method to Reduce The Effect of Thermal Drift in Optical Resonant Sensors 54 4.1 Introduction 54 4.2 Theoretical Principle 55 4.3 Experimental Demonstration 63 vi 4.4 Conclusion 75 4.5 Chapter References 76 Chapter 5: Integrated Microfluidic/ Whispering Gallery Mode Based Sensor Systems for Chemical and Biological Analysis 78 5.1 Microfluidics and Its Advantages 78 5.2 Ultra High Q WGM Based Sensors and The Need of Practical Devices 80 5.3 Fabrication a Prototype of The Integration High Q WGM Resonator with Microfluidics Channel (Opto-Fluidics Devices) 81 5.4 Experimental Demonstration and Results 85 5.5 Chapter References 90 Chapter 6: Reaction Kinetics Simulation of Microfluidics Whispering Gallery Mode Resonator Based Sensors 91 6.1 Introduction 91 6.2 Theoretical Analysis of Microfluidic Assisted WGM Based Sensors 92 6.2.1 Mass-Transport (Convection-Diffusion) Process 94 6.2.2 Chemical Surface Reaction 98 6.3 Computational Simulations 102 6.3.1 Effects of Flow Rate on The Reaction Kinetics 104 6.3.2 Effects of Channel Height on The Reaction Kinetics 107 6.3.3 Effects of Bulk Concentration 108 6.4 Discussions and Summary 109 6.5 Chapter References 111 Chapter 7: Practical Applications: Integrated, Array and Portable Devices 113 7.1 Real-Time Ultra Sensitive Sensing System for Influenza A Virus Detection 113 7.1.1 Introduction 113 7.1.2 Theoretical Principle 115 7.1.3 Fabrication 119 7.2 Array WGM Based Sensors for Multiple Biomolecules Targets Detection 121 7.3 Ultimate Sensitivity, Portable and On-Site Detection Whispering Gallery Mode Based Sensors 124 7.3.1 Theory Operation and Design Procedure of Electronics Readout Circuits 125 7.3.2 Circuit Implementation and Experimental Demonstration of a Practical Portable Sensor 127 7.4 Chapter References 134 Chapter 8: Conclusions and Future Research Development 136 Bibliography 140 vii List of Tables Table 6.1: The association rate constant, the disassociation rate constant and the binding affinity of various type of analytes. 101 Table 7.1: The Molecular Weight and the Mass of a single molecule or virus. 118 viii List of Figures Figure 2.1: A plan view image of St. Paul’s Cathedral in London where Lord Rayleigh observed acoustic waves clinging to the dome of the building. 8 Figure 2.2: Geometric Schematic of the WGM resonance. 14 Figure 2.3: Disk resonator in cylindrical coordination and its quantum mode numbers. 15 Figure 2.4: The graph is a field intensity profile of the TE mode in a radial direction for the fused silica disk resonator with radius of 1.6 mm. The wavelength is at 1550nm and the mode numbers are q = 1 (Upper) and q = 2 (Lower) for l = 9327. 18 Figure 2.5: Field distribution of the lowest mode in z direction of a WGM resonator disk with the thickness of 50 µm. 19 Figure 2.6: Intensity distribution of Whispering Gallery fundamental mode with q = 1 and l = m = 15. The resonator boundary is indicated by a black circle, where the decaying evanescent field outside of the resonator can be observed. 20 Figure 2.7: Field distribution of Fused Silica WGM resonator disk d = 3.2 mm, t = 50 µm. Upper: Fundamental TM mode. Lower: High order TM mode with l – m = 1. 21 Figure 2.8: Scattering losses from residual surface in homogeneities ss Q , and the absorption losses due to a contamination on the surface of the resonator w Q as a function of resonator’s diameter. 26 Figure 2.9: Free Spectrum Range as a function of the Fused Silica resonator’s radius at a wavelength of 1550 nm. 27 Figure 3.1: Optical polymer waveguide coupled to high Q WGM resonator. 33 Figure 3.2: Normalized Transmission Output in different coupling regimes: under- coupled regime (green), critical coupling (red), and over-coupled regime (blue). 36 ix Figure 3.3: Precision Polishing TechPrep System Model #15-2000 from an Allied High Tech Product, Inc. 37 Figure 3.4: High Q WGM Fused Silica resonators. 38 Figure 3.5: The plot is the effective refractive index of WG fundamental modes TE (red) and TM (blue) as a function of the resonator’s radius, with resonator’s material as Fused Silica with refractive index of n = 1.444, at wavelength λ = 1550 nm, and the surrounding medium is air. 38 Figure 3.6: The field intensity of the polymer waveguide was calculated by the Effective Index Method. 40 Figure 3.7: The fabrication procedure and dimensions of a polymer waveguide. 41 Figure 3.8: The Field Intensity Distribution of the polymer waveguide was simulated by the Multi-physics Modeling and Simulation Software from COMSOL, with the Finite Element Method (FEM). 42 Figure 3.9: Optical polymer waveguide with the fiber butt coupled by a single mode Fiber. Upper: SMF was butt-coupled into one of two polymer waveguides in the same substrate. Lower: Red laser 634nm was launched into polymer waveguide by fiber butt-coupling. 43 Figure 3.10: Overview illustration of Polymer waveguide vertically coupled to high Q WGM resonators. 44 Figure 3.11: Experimental demonstration of optical Polymer waveguide vertically coupled to high Q WGM Fused Silica resonator disk. 45 Figure 3.12: The optical transmission spectrum for the polymer waveguide was vertically coupled to the high Q WGM Fused Silica Resonator disk at 1550nm: (Upper) The output power versus laser detuning frequency. (Lower) The expanded frequency scale is showing a Loaded Q of 1.2 x 10 7 . 47 Figure 3.13: Transmission factor versus the gap between the optical polymer waveguide and the WGM resonator disk. 48 Figure 3.14: Integration of the polymer waveguide, vertically coupled to the high Q WGM resonator in a practical packet. 50 Figure 3.15: Integration of the polymer waveguide, vertically coupled to the high Q WGM resonator: Normalized Transmission Output Intensity versus Laser Detuning frequency. 51 x Figure 4.1: The plot illustrates resonance frequencies shifting due to a temperature fluctuation for an individual TE or TM mode and the change of the differential frequency of the two resonance modes. The theoretical calculation is based on 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, angular mode l = 9331, radial mode q = 1. Upper: The case of surrounding medium is air (n 0 = 1). Lower: The case of surrounding medium is water (n 0 = 1.333). 59 Figure 4.2: The plot illustrates resonance frequencies shifting due to the change in the refractive index of the surrounding medium for an individual TE or TM mode and the change of the differential frequency of two resonance modes. The theoretical calculation is based on a 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, for the case of angular mode l = 9331, radial mode q = 1. The surrounding medium is water (n 0 = 1.333). 61 Figure 4.3: The plot illustrates resonance frequencies shift due to the change in refractive index of surrounding medium for individual TE or TM mode and the change of differential frequency of two resonance modes. The theoretical calculation is based on 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, for the case of angular mode l = 9331, radial mode q = 1. The surrounding medium is air (n 0 = 1), and the lower plot is the expanding calculation of the change of outside refractive index to 10 -8 RIU. 62 Figure 4.4: Experiment setup. 63 Figure 4.5: Thorlabs’ Manual Fiber Polarization Controller FPC560. 64 Figure 4.6: The Transmission Output Spectrum is adjusted by the positions of the Thorlabs Manual Fiber Polarization Controller to achieve different polarization coupling states to the resonator. Upper: TM mode was strongly coupled to WGM resonator. Lower: TE mode was strongly coupled to WGM resonator. 65 Figure 4.7: The Transmission Output Spectrum is adjusted by the positions of the Thorlabs Manual Fiber Polarization Controller to achieve different polarization coupling states to the resonator. Both TE and TM modes were coupled to WGM resonator by control the position of Polarization Controller. 66 Figure 4.8: The magnified view of a Transmission Output Spectrum to show a flat portion at the resonant peak area. 67 xi Figure 4.9: Upper: Output Transmission spectrum. Lower: A Lock-in Amplifier can obtain the derivative of the transmission output signal and will help to measure the shift of resonant peaks by tracking its zero crossing points. 69 Figure 4.10: The plot represents theoretical calculation vs. experimental data for the resonance shift of each individual TE and TM modes and the change of the differential frequencies when the temperature is changed by 5 o C (the experimental system consists of Fused Silica resonator surrounded by air). Crosses, dots, and triangles are corresponding to TE, TM resonance shifts and the differences, respectively. The dash, dot, and solid curves are the theory prediction of individual TM and TE resonant frequency shifts and its difference. The uncertainty of TM or TE resonance mode shift is ± 0.65pm. The observed changes in the differential frequency were within the accuracy of our measurement setup (± 0.05pm). Theoretically, the differential resonance shift is about 0.001 pm when temperature is changed by 5 o C. 71 Figure 4.11: The plot illustrates the experimental data and the theory prediction. The changes in the outside refractive index were calculated for the three Glucose solutions. 73 Figure 4.12: The output transmission spectrum, when the detected samples are DI water and Glucose concentrations of 0.5 wt% and 1 wt%. 74 Figure 5.1: The construction of the Microfluidics WGM based sensing cell. 82 Figure 5.2: High Q WGM resonator (fabricated with 2 spacers on the top and the bottom surfaces) integrated with microfluidics. 83 Figure 5.3: High Q WGM resonator (fabricated by polishing a cone shape at the edge of a resonator) integrated with microfluidics. 84 Figure 5.4: High Q WGM based microfluidics sensing: an experimental setup. 86 Figure 5.5: Experimental data: the resonance wavelength shifts for detecting 1 wt.%, 1.5 wt.%, 2 wt.%, 2.5 wt.% and 3 wt.% Glucose solutions and the theory prediction. The changes in the outside refractive index were calculated for glucose solutions and by using the differential TE and TM mode measurement method. 88 Figure 5.6: The sample volume vs. the radius of WGM resonators. 89 xii Figure 6.1: A study of the model of the sensing system. The solution of the target analyte with concentration c 0 flows with velocity v, through our microfluidics channel of height h, over our sensor based on the WGM resonator. The edge surface of the resonator is functionalized with b m receptors per unit area (mol/m 2 ). The polymer waveguide is used to couple the light in and out and it can be glued with the microfluidics channel to form a sensing cell with one inlet and two outlets. 93 Figure 6.2: The general model of the microfluidics sensors which involves convection, diffusion, and surface reaction in a flat geometry. The microfluidics channel with height h, and the sensor’s surface with length L s . The bulk analyte is affected by the convection and the diffusion processes to diffuse to the reaction to the surface, where the binding process will happen. 94 Figure 6.3: The diffusion time as a function of the analyte’s distance from the sensor reaction surface. 95 Figure 6.4: The depletion layer thickness as a function of flow velocity of bulk concentration through the microfluidics channel when P es >> 1. 98 Figure 6.5: The saturated concentration of antibody-antigen complex as a function of the target concentration c s at the reaction surface. The red curve for Binding Affinity K D = 7.5*10 -14 M and the blue curve for K D = 10 -9 M. 100 Figure 6.6: The ideal reaction time for different antibodies plotted against their analyte concentration. The simulation was calculated with affinity parameters as shown in the table below. The red curve is for Streptavidin/ Biotin; the blue curve is for Human C-reactive Protein; the black curve is for Immunoglobulin G; and the pink curve is for Protein A33 Immunoassay. 100 Figure 6.7: Calculation of Damkohler number based on the thickness of depletion layer that is a function of the flow velocity. Red curve is for k a = 4 x 10 4 (m 3 /(mol.s)) and blue is for k a = 10 3 (m 3 /(mol.s)). Both curves used b m = 4*10 -8 mol/m 2 and D = 7.4*10 -11 m 2 /s. 102 Figure 6.8: A vertical cross section of the flow field through the microfluidics channel, incorporated with WGM resonator. 104 Figure 6.9: Upper: The plot of surface concentration b(t) of receptors that are bound by target molecules as a function of total time which obtained by COMSOL simulation. Lower: The total time which includes all processes (convection, diffusion and surface reaction) as a function of flow velocity. We simulated with different flow velocities: 2µm/s, 5µm/s, 10µm/s, 20µm/s, 50µm/s and 100µm/s. 106 xiii Figure 6.10: The total detection time with different channel heights: 50µm, 100µm, 150µm, 200µm, 400µm and 1mm which obtained by COMSOL simulation. 108 Figure 6.11: The total detection time with the different analyte concentrations: 10nM, 50nM, 100nM, 200nM, and 500nM, which is obtained by the COMSOL simulation. 109 Figure 7.1: Antibody-AntigenInteraction: Illustrated by the antibody (Y shape), the antibody's specific target (antigen; diamond shape), and an unspecific target (circle shape). 115 Figure 7.2: The theoretical plot of the Differential TE and TM modes resonant wavelength shift, as a function of the mass of the molecular layer on a detecting surface of sensors. The calculation is based on Fused Silica WGM resonator with radius of 1.6 mm. 117 Figure 7.3: The proposed integrated chip of the optical sensor based on a polymer optical waveguide vertically coupled to the high Q resonator disk. 120 Figure 7.4: Proposed series and parallel platforms array sensors for multiple targets detection 123 Figure 7.5: Functionalized Optical Microresonators in an array would address optically via row and column optical polymer waveguides. It is a proposed combination platform series and parallel detection array sensors. 124 Figure 7.6: Schematics of our proposed microfluidics WGM based portable sensor. 126 Figure 7.7: Schematics of the electronics circuit. 127 Figure 7.8: Function blocks of the electronics circuit. 130 Figure 7.9: The signals captured by the oscilloscope show the “START” and “STOP” control points. The square waves are counted by the counter between the time interval of two TE and TM mode resonant peaks, as the derivative of output transmission comes from output of the Lock in Amplifier. 131 Figure 7.10: The electronics module and the real time measurement display of the total number of rising edges that were counted by the counter. They are proportional to the differential frequency between the TE and TM resonant peaks. The lower picture is the electronics circuit inside the module. 132 xiv Abstract This work is a study of a practical Whispering Gallery Mode based sensing device that consists of optical polymer waveguides integrated with high Q Whispering Gallery Mode (WGM) resonators that is incorporated with microfluidics technology. We proposed, developed and experimental demonstrated a method that solves the issue of integrating high Q WGM resonators into practical sensors for real time and portable on field detection. We demonstrated a novel implementation of a polymer optical waveguide coupled to the whispering gallery modes of a fused silica resonator disk. This is the first step in bringing high Q WGM disks closer to the integrated optics technology. We obtained near-critical coupling and a loaded Q = 1.2 x 10 7 . The use of a WGM resonator with an ultra-high quality factor Q is promising in highly sensitive, label-free, lab-on-a-chip sensor applications. However, one issue that causes the reduction of sensitivity and detection limits of the resonant sensor class is the fluctuation of temperature in the surrounding medium. We investigated a novel method of using the differential frequency of TE and TM modes to reduce the thermal noise baseline. Then, we studied the temperature dependence of the WGM based sensors and experimentally demonstrated the reduction of temperature fluctuation; and, thus, there was a significant improvement in the practical sensor detection limit. Consequently, the United States Patent 8,111,402 was granted for our novel invention. In additional, the optical sensors are desirable to analyze small volumes or extremely dilute concentrations of analyte, thus we need an efficient method of delivering xv target molecules to the sensor’s surface. Microfluidics and the concept of micro total analysis systems (µTAS) is a new and promising technology expected to revolutionize chemical and medical analysis systems. We theoretical analyzed and experimental demonstrated an integrating prototype of polymer waveguide-high Q WGM based sensor with the microfluidics technology. It is the critical step to bringing the prototype into a practical biosensor, which will lead to the ability to form a so-called lab-on-a-chip system. 1 Chapter 1 Introduction 1.1 Background and Motivation Recently, there has been great interest for research in creating an efficient practical sensor device, in order to detect chemical or biological molecules in various different fields such as environmental control, bacterial and virus detection, drug evaluation, military applications, and even in homeland security. Techniques such as spectroscopy are valuable sensing methods in a laboratory environment; however, it is time consuming, expensive, and requires bulky instruments. Outside the laboratory, there is a significant need for robust and inexpensive sensors offering a real-time and on-field detection measurement without the necessity of a complicated sample preparation. A compact, label free, high sensitivity, and selective sensing system are, of course, all important prerequisites. In the evolution of optical sensing technology, one can measure nearly all of the physical measurands of interest and a very large number of chemical quantities- such as: temperature, pressure, rotation, acceleration, humidity, hydrogen, biochemical, and many more. During recent years, research for development optical sensors are motivated by their significant advantages such as high sensitivity, electrical passiveness, freedom from electromagnetic interference, no power consumption, and multiplexing capabilities. For example, an optical hydrogen sensor is desirable in a potentially explosive environment, where it requires no electrical spark or optical sensor network, which utilizes the 2 capability of fiber optic for sending and receiving optical signals over long distances, avoid converting between electronics and photonics at each sensing site; therefore, it can reduce the costs and increase the flexibility. In addition, measurements using light as a means of detection are essentially the fastest in response because of their inherent propagation velocity, although the response time is limited by the electronics detection. A modern sensor system incorporates several key components such as a sample delivery unit, sensing material, transducer, and data processor. Developing sensing system is a recognized challenge because it requires extensive experimentation not only to achieve the best performance, but also stability, manufacturability, cost, and other practical issues. Optical biosensors are typically transducers that detect the presence of molecules, chemical or biological at their surface. Their advantages and desirable features include its extreme sensitivity (nM or less), which is non-destructive to the sample and the processes of detection generally on its surface, so can be tailored to sense almost any kind of molecules. Whispering gallery modes resonators in various forms of sphere, disk or ring are known to have extremely high quality factors (Q) in the optical domain. The ultra smooth surface and very low loss of resonators’ materials lead to its ultrahigh Quality factor, which was measured to be as high as 10 10 in the infrared for a silica microsphere [1, 2]. The ultrahigh Q, in long light storage time, gives very high circulating power in the cavity. Therefore, Whispering Gallery Mode resonators represent a new sensing mechanism; thus, they have recently raised attention in bio/chemical sensor development. They proved that they are among one of the most sensitive class of biosensor. The light is 3 trapped near the edge of WGM resonators by the total refection, produces an evanescent field outside the resonators with a characteristic length of approximately 100nm (depending on the relative index of refraction of the waveguide and the sample medium), and is sensitive to the refractive index change induced by the binding (or removal) of bio/chemical molecules to and from the resonators’ surface. Surface treatments, such as antibodies or oligonucleotide strands, can provide specificity for the analyte in the detection purposes. The motivations behind this work are that these molecules are traditionally detected with fluorescence approaches that involve tedious labeling processes [3, 4]. WGM resonator based sensors also offer a sensitive label-free detection alternative with multiplexing capability. Unlike the conventional waveguide-based sensors in which the light passes through the waveguide only once, the WGM circulates repeatedly around the resonators surface. This recycling of the light significantly enhances the interaction between the WGM and the analytes on the surrounding medium near the resonator’s surface. The enhancement can be characterized by the Q-factor, which is typically 10 6 to 10 9 for a fused silica resonator. Despite the small physical size of a resonator, such high Q-factors result in an effective interaction length on the order of 10 cm to 100 cm that in turn leads to a significant improved sensitivity and a lower detection limit. In additional, optical polymer waveguides play an important role in rapidly developing areas such as communication, optical networking, and photonics. Their combined advantages also make them an ideal integration platform with various materials, such as YIG, LiNbO3, and as a semiconductor to enable constructing of an 4 optical device (amplifier modules or optical add/drop multiplexer) in a single substrate. Moreover, the flexibility and toughness in optical polymer materials pose an intensive ability for 3D integration or even all-polymer integrated optics. State-of-the-art polymer also offers rapid processibility, cost effectiveness, and mature fabrication methods; thus, it is suitable to incorporate with high Q WGM resonators (often made by Fused Silica, CaF2 or MgF 2 ) for a highly integrated optical platform. Tapered fibers, angle polished fibers, and frustrated total internal reflection prisms are typically used to couple to the whispering gallery modes in these resonators. However, integration of all sensing device into a photonics device is still a challenge that includes these problems: alignment or methods of mechanically stabling the platform. On the other hand, ring waveguide and disk micro-resonators have been integrated with coupling bus waveguides in several materials systems [5-7], but the highest reported Q’s in these resonators are ~10 5 , usually limited by scattering from sidewalls, which is reducing the sensitivity of the device we want to built. The motivation to incorporate the high Q fused silica disks with the polymer integrated circuit technology is a result of trying to resolve these concerns. This technology will widely open the ability to develop an integrated photonics device that can sense an array of implementation for parallel detection and on-field portability. One of the most significant common problems of the high Q WGM resonators is that the thermal fluctuations of, either, the ambient environmental temperature variations or the absorption of laser energy during the laser scanning or pumping will affect the performance of the device. For example, the thermal drift of a WGM resonant frequency, 5 caused by the laser energy absorption during a laser scanning or the fluctuation of surrounding medium temperatures, could easily exceed 100 times the resonant line width when the Q factor of resonators is higher than 10 8 [8]. This thermal instability of the resonant frequency in a resonator causes a wide range of undesirable thermal-induced phenomena such as hysteretic wavelength response, oscillatory instability [8, 9] or a dramatic reduction in the sensitivity of the sensing devices. As the motivation for my theory and experimental demonstration, the problem influenced me to compensate the thermal drift of the resonant frequency in our sensing devices by the differential of TE and TM mode measurement. A practical operation of the sensors, in a complex microfluidic environment, demands for a simple and smart way for the excitation and detection of the WGM. For this reason, we proposed a preliminary demonstration of the simplest conceivable scheme of integrating the high Q WGM based sensing system into the microfluidics technology. 1.2 Dissertation Outline In this thesis, the ultra high Q WGM resonators, coupled with optical polymer waveguides, are investigated and analyzed for the first time. We performed both theoretical analysis and experimental demonstration on a practical sensing system consisting of polymer waveguides coupled with high Q WGM resonators, incorporating it to microfluidics technology to produce a compact, high sensitivity, portable, real-time, on-field detection. 6 Starting in Chapter 2, it is a basic introduction to WGM resonators disk. The resonant characteristics of a resonator, such as their field distributions or quality factor, are discussed and serve as an introduction to the terminology that is used throughout this thesis. Next, Chapter 3 is our theoretical design calculation and experimental demonstration of optical polymer waveguides coupled with high Q WGM resonators. Coupling between a waveguide and any cavity relies upon the physical overlap and vector phase matching of the evanescent field components of the waveguide and the resonator disk. They must be phase matched to allow efficient optical coupling from one to the other. Therefore, the effective refractive indices of the waveguide and resonator need to be designed and fabricated to be almost equal. In Chapter 4, we investigated and demonstrated a novel method of differentiating between the TE and TM mode resonant frequencies measurement for significantly reducing the effect of thermal drift in WGM resonator based sensors, without the need of temperature stabilization. The technique leads to a practically achievable low detection limit in the presence of temperature changes. Then, Chapter 5 describes our work on integrating the microfluidics technology into our WGM based sensing system for chemical and biological detection and analysis. The fabrication processes and experimental demonstration were reported to prove our concept. Thus, Chapter 6 is the reaction kinetics simulation of microfluidics WGM resonator based sensors, which leads to Chapter 7, which is our proposal of practical applications toward a portable integrated array device for detecting multiple bimolecular targets. Finally, Chapter 8 is the conclusion of our thesis and future research and development. 7 1.3 Chapter References [1] V. B. Braginsky, M. L. Gorodetsky, and V. S. Ilchenko, “Quality-factor and non- linear properties of optical whispering-gallery modes,” Physics Letters A, vol.137, pp. 393-397, 1989. [2] L. Collot, V. Lefèvre-Seguin, M. Brune, J. M. Raimond, and S. Haroche, “Very High-Q Whispering-Gallery Mode Resonances Observed on Fused Silica Microspheres,” Europhysics Letters, vol. 23, no. 5, pp. 327-334, 1993. [3] T. Vo-Dinh, L. Allain, “Biosensors for Medical Applications”, in Biomedical Photonics Handbook, T. Vo-Dinh, Ed. CRC Press, 2003. [4] T. Vo-Dinh, B. M. Cullum, “Fluorescence Spectroscopy for Biomedical Diagnostics”, in Biomedical Photonics Handbook, T. Vo-Dinh, Ed. CRC Press, 2003. [5] B. E. Little, S. T. Chu, W. Pan, D. Ripin, T. Kaneko, Y. Kokubun, and E. Ippen, “Vertically coupled glass microring resonator channel dropping filters,” IEEE Photonics Technology Letters, vol. 11, no. 2, pp. 215-217, Feb. 1999. [6] P. Rabiei, W. H. Steier, C. Zhang, and L. R. Dalton, “Polymer micro ring filters and modulators,” Journal of Lightwave Technology, vol. 20, no. 11, pp. 1968-1975, Nov. 2002. [7] D. V. Tishinin, P. D. Dapkus, A. E. Bond, I. Kim, C. K. Lin, and J. O’Brien, “Vertical resonant couplers with precision coupling efficiency control fabricated by wafer bonding,” IEEE Photonics Technology Letters, vol. 11, no. 8, pp. 1003-1005, Aug. 1999. [8] T. Carmon, L. Yang, and K. J. Vahala, “Dynamical thermal behavior and thermal self-stability of microcavities,” Optics Express, vol. 12, no. 20, pp. 4742-4750, 2004. [9] A. A. Savchenkov, V. S. Ilchenko, A. B. Matsko, and L. Maleki, “Kilohertz optical resonances in dielectric crystal cavities,” Physical Review A, vol. 70, no. 5, 2004. 8 Chapter 2 Whispering Gallery Mode Resonators 2.1 Introduction The whispering gallery phenomena was first introduced by Lord Rayleigh in 1910 [1-3] when he described the phenomenon of acoustical waves. He observed these while the waves were propagating around the interior gallery of the Saint Paul’s Cathedral in London. He noticed the acoustical modes were partially confined due to the suppression of the wave diffraction by the sound reflection from the curved dome walls. Figure 2.1: A plan view image of St. Paul’s Cathedral in London where Lord Rayleigh observed acoustic waves clinging to the dome of the building. 9 This kind of resonance phenomenon is also true with light waves. An optical cavity with circular geometry is called whispering gallery mode resonator. Since WGM is a morphology-dependent phenomenon, the radius of the resonator determines the effective volumes and distribution of the modes. Thus, the optical WGM cavity is investigated in many different shapes and was significantly shrank down in scale to a very small size in millimeter, or even to nanometer, in order to be suitable for demonstration in many applications. At the beginning of the 20 th century, Mie was the first to investigate the optical modes of a spherical dielectric micro particle, within the context of light scattering from spherical particles. The sharp features of the scattering spectrum were attributed to the resonant circulation of the optical energy within the optical cavity. Light injected into the WGM resonators is trapped in circular orbits within the surface by repeated total internal reflections. First recognized by Braginsky, Gorodetsky and Ilchenko in their pioneering study of ultrahigh Q microspheres and predicted numerous nonlinear applications [4], the surface tension induced solid WGM silica microspheres, which posed the low dissipation mechanism of confinement, resulted in an unusually high quality factor. As a result, Savchenkov et al. [5] at Jet Propulsion Lab (JPL) in Pasadena, CA have demonstrated a millimeter sized CaF 2 WGM resonator with ultrahigh quality factor (Q) higher than 10 11 . This process of fabrication is relatively simple and inexpensive. A wide range of resonators built for different applications have been fabricated and investigated in the last two decades. Optical WGM have been realized in high quality dielectric resonators of many different shapes such as cylindrical [6], spherical [4], 10 toroidal [7], ring [8, 9], and etc. They have drawn increasing attention due to their great potential in the application of cavity quantum electrodynamics [10], low threshold and narrow line-‐width lasers [11, 12], optical filters [9], high-‐sensitivity transducers for miniature sensors [13, 14], and etc. There are many important highlights in their applications and significant achievements. The solid-state WGM lasers, which was studied by Garrett et al. [15] with CaF2:Sm ++ crystalline resonators, can be considered as the first contribution of WGM in optics. The resonator was a millimeter sphere with highly polished surfaces. During which, a pulsed laser operation, due to a total internal reflection in a ruby ring at room temperature, was observed. From then, the WGM based lasers became one of the most obvious applications in optical domain; and its advantage – the high quality factor of the resonators – was a significant help in achieving narrow linewidth and reducing the laser threshold. Knight et al. [16] demonstrated laser emission from WGM in a highly refractive dye-doped solvent flowing in a normally illuminated silica capillary. Another way to create a WGM laser is by the use of solids doped with rare earth ions. Sandoghdar et al. [11] realized a WGM laser based on neodymium-doped silica microspheres with a 200-nW threshold. Cai et al. [12] demonstrated heavily doped Er:Yb phosphate glass microsphere, with large gain, which coupled to fiber taper for generating laser at 1.5µm with an output power as high as 3µW, single mode operation and achieved a laser threshold pump power of 60µW. Resonators are fabricated with active electro-optic material such as LiNbO3, which can be tunable and combine with their narrow linewidth property, leads to the 11 optical filter research. The narrow linewidth will benefit for channel capacity increasing and the tunability will determine the overall signal quality, sensitivity, and efficiency. Ilchenko et al. [17] demonstrated a high Q microwave filter with a linewidth of about 10MHz and a tuning range exceeding 10GHz, utilizing a WGM resonator fabricated with LiNbO 3 . Savchenkov et al. [18] have produced optical filters with a bandwidth of about 10 KHz using CaF 2 WGM resonators, which can be tuned by temperature change. The insertion loss of the filter was at 5-dB level. In recent years, WGM resonators have attracted more attention and been under intensive research in the field of chemical and biological sensing due to their very high quality factor, small volume of sample requirement, and being label free. The WGM is produced by the total internal reflection of light that is confined inside the resonator, close to its surface, and has the evanescent field extending into the surrounding medium. Thus, this enables the interaction with the analyte near the resonator’s surface. The light in WGM resonators can orbit many thousand times before escaping; therefore, the detection sensitivity of this method is expected to be greatly enhanced. The presence of chemical or biological molecules and its concentration on the resonator’s surface can then be detected by measuring the frequency shift of a WGM caused by the analyte’s perturbation of the index of refraction [19, 20]. In addition, analyte absorption will change the effective quality factor Q of the WGM; and the modification of mode linewidth can be measured [21]. We can also envision the use of the cavity-ringdown spectroscopy [22]. Surface treatments such as antibodies immobilization can provide specificity for the analyte so that the sensor only detects selective target molecules. For label-free 12 biomolecular detection, Arnold et al. [23] have studied a sensing platform consisting of a silica microsphere coupled with a fiber taper. Whenever chemical or biological molecules are attached on the surface of the microsphere, it results in the changes of the effective resonator diameter and refractive index, which causes a shift of the WGM resonance frequency, effecting detection. Based on a perturbation analysis, they predicted that the shift should be inversely proportional to the microsphere radius, proportional to protein surface density, and access polarization [19]. Fan’s group at the University of Missouri introduced a liquid-core optical ring- resonator sensor (LCORRS). It is a silica capillary coupled to the waist of an optical fiber taper. The light propagating along the taper waist excites the whispering gallery modes (WGMs) inside the capillary wall. The capillary wall has a thickness of a few microns. For this reason, the evanescent part of the WGMs can penetrate into the interior part of the capillary and is able to probe the liquid flowing inside the capillary. They demonstrated in a variety of sensing applications such as: chemical vapor, DNT, viral, and bimolecular detections [24]. In 2007, Andrea M. Armani et al. [25] achieved an ultimate sensitivity result by using a silica micro toroid to employ light recirculation. They easily reached a single molecule sensitivity by increasing the wavelength shift with thermo-optic boost mechanism. The amplification of the shift was large enough to demonstrate the detection of an individual Interleukin-2 (IL-2) protein from the signal generated by cytokines, binding to antibodies that were previously immobilized on the toroid surface. The reported sensitivity 10 -16 M IL-2 for the thermo-optic WGM technique exceeded that of 13 even the most sensitive enzyme-linked immunosorbent assay (ELISA). They also demonstrated a dynamic range of 10 12 in concentration, establishing the WGM micro resonator as a sensitive and versatile detector. In this chapter, the WGM resonant characteristics of the dielectric disk resonator, such as field distribution, cavity quality factor, loss mechanisms, free spectral range, and finesse are discussed and serve as an introduction to the terminology which is used throughout this thesis. 2.2 Geometric Optics Analysis of WGM Resonators Consider a resonator disk, which has radius R. The lights are confined inside the resonator by a total internal reflection (TIR) because the angels between beams and normal directions of the boundary satisfy Snell’s law. The propagation path of the light inside the resonator can be approximated as a polygon, as shown in Figure 2.2, and the smaller the resonance wavelength, in the higher order of the polygon. The resonators usually satisfied the condition of R >> λ 0, so that the distance of travelling path of photons inside the optical disk resonator is equal to its equatorial circumference. By applying the phase-matching principle of the resonance, we have 2!"! !"" = !! ! (2.1) where n eff is the effective refractive index of the resonator, R is its radius, λ 0 is the resonance wavelength, and l is the angular mode number which corresponds to the number of wavelengths along the resonator surface 2!Rn eff . 14 Figure 2.2: Geometric Schematic of the WGM resonance. 2.3 Analytic Approximation for The Modes in a WGM Disk Resonator In Maxwell’s equation, an isotropic medium, with constant scalar permittivity and permeability, is free of charge and current to investigate the electromagnetic fields in disk resonators. We have E i H H i E ωε ωµ = × ∇ − = × ∇ (2.2) First, we took the curl of Maxwell’s curl equation for E field and used the vector identity E H i E ) E ( E µε ω ωµ 2 2 = × ∇ − = ∇ − ⋅ ∇ ∇ = × ∇ × ∇ (2.3) Then, we obtain 0 2 2 = − ∇ E E β (2.4) where µε ω β = R 15 Figure 2.3: Disk resonator in cylindrical coordination; and its quantum mode numbers. In a cylindrical coordinate, a general solution to the Helmholtz’s equation for source free and lossless media can be written as ) z , , ( E a ˆ ) z , , ( E a ˆ ) z , , ( E a ˆ ) z , , ( E z z φ ρ + φ ρ + φ ρ = φ ρ φ φ ρ ρ (2.5) Thus, we can expand wave equation to three scalar partial differential equations of the form: ρ φ ρ ρ β − = φ ∂ ∂ ρ − ρ − + ∇ E ) E E ( E 2 2 2 2 2 φ ρ φ φ β − = φ ∂ ∂ ρ − ρ − + ∇ E ) E E ( E 2 2 2 2 2 (2.6) z z E E 2 2 β − = ∇ After applying the Laplacian of a scalar in cylindrical coordinates, we get F z F F 1 F 1 F F 2 2 2 2 2 2 2 2 β φ ρ ρ ρ ρ − = ∂ ∂ + ∂ ∂ + ∂ ∂ + ∂ ∂ = ∇ (2.7) where F = {E,H}. m l 16 In the disk resonator, the confinement in the vertical direction restricts the movement of a photon in a plane. In this case there are two polarizations: TE (E field parallel to the disk plane) and TM (E field perpendicular to the disk plane), and the equation (2.6) becomes scalar in the z direction. Thus, z F corresponds to z H ( z E ) for TE (TM) modes. Assume separable solution for ) z , , ( F z φ ρ of the form ) z ( ) ( ) ( ) z , , ( F z Ζ Φ Ψ φ ρ = φ ρ (2.8) We get 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 β Ζ φ Φ ρ Φ ρ Ψ ρ Ψ ρ Ψ Ψ ΨΦΖ β ΨΦ φ Φ ρ ΨΖ ρ Ψ ρ ΦΖ ρ Ψ ΦΖ − = ∂ ∂ + ∂ ∂ + ∂ ∂ + ∂ ∂ − = ∂ ∂ + ∂ ∂ + ∂ ∂ + ∂ ∂ z Z z Z (2.9) Let Z z Z z 2 2 2 β − = ∂ ∂ (2.10) Φ Φ 2 2 2 l − = φ ∂ ∂ (2.11) and 2 2 2 β = β + β ρ z (2.12) Consequently, it leads to a classic Bessel Differential Equation: ! 2 ! 2 " !! 2 +! !" !! + (" ! !) 2 #l 2 $ % & ' "=0 (2.13) 17 These solutions are called spherical Bessel functions of order l. Within the disk, the general solution is ) ( BY ) ( AJ l l ρ β + ρ β = ρ ρ Ψ (2.14) ) ( J l ρ β ρ and ) ( Y l ρ β ρ represent, respectively, the Bessel function of the first and second kind, in which l is the angular mode number. We know that ) ( Y l ρ β ρ is divergent as 0 → ρ . Since the field should be finite at R = 0, then the solution will become ) ( AJ l ρ β = ρ Ψ for ρ ≤ R (2.15) and A is constant and can be determined by boundary conditions. Outside of the disk it is an evanescent field that decay exponentially in the radial direction )) R ( exp( ) R ( J l − ρ ⋅ α − ⋅ β = ρ Ψ for ρ > R (2.16) with 2 2 2 o eff n n − λ π = α The solution of the equation (2.10) follows the standard slab mode calculation [27]: ) z cos( ) z ( Z z β = (2.17) where t m z π = β and t is the thickness of disk resonator and m is longitudinal mode number. 18 Figure 2.4: The graph is the field intensity profile of the TE mode in the radial direction for a fused silica disk resonator with radius of 1.6 mm. The wavelength is at 1550nm and the mode numbers are q = 1 (upper) and q = 2 (Lower) for l = 9327. 19 40 − 20 − 0 20 40 0 0.2 0.4 0.6 0.8 1 Position (um) Normalized Field Amplitude Figure 2.5: Field distribution of lowest mode in z direction of WGM resonator disk with thickness of 50 µm. 24 25 26 27 0 0.02 0.04 0.06 0.08 Position (um) Normalized Field Amplitude Surface of resonator Evanescent Field 20 The lowest mode in z direction is calculated as in figure 2.5 for the resonator disk thickness of 50 µm. We can see that an evanescent field can be observed outside of a resonator as shown in figure 2.5. While the solution of equation (2.11) is simply ) (φ Φ ~ ) il exp( φ (2.18) The intensity distribution of a fundamental WGM mode in a resonator can be calculated and plotted by using the MATHLAB code which is based on the above analysis equations. A top view of a fundamental mode (q = 1 and l = m = 15) in a resonator disk can be shown as in figure 2.6. There are 30 maxima, corresponding to l = m = 15. The evanescent field of the mode is visible outside of the resonator boundary. Figure 2.6: Intensity distribution of Whispering Gallery fundamental mode with q = 1 and l = m = 15. The resonator boundary is indicated by black circle, and the decaying evanescent field outside of the resonator can be observed. 21 Figure 2.7: Field distribution of Fused Silica WGM resonator disk d = 3.2 mm, t = 50 µm Upper: Fundamental TM mode. Lower: High order TM mode with l – m = 1 Besides the analytic solutions, we have also modeled the WGM resonator disk through a computer-aided design (CAD) program in conjunction with a program that numerically solves Maxwell’s equations. Additionally, we used a popular commercial software package called Multiphysics Modeling and Simulation Software from COMSOL for efficiently calculating the frequencies and fields of the Whispering Gallery 22 modes of the axisymmetric dielectric resonator disks [27]. In Figure 2.7, the field distribution results from our simulation with a Fused Silica disk resonator with a diameter of 3.2 mm and a thickness of 50 µm at a wavelength around 1550nm. 2.4 Properties of WGM Resonator 2.4.1 Quality Factor Q and System Loss The loss of a resonator is an important parameter and it is used to determine the potential pertinence of a resonator for many different applications. Resonator loss is commonly expressed in term of the quality factor Q. The quality factor Q is related both to the linewidth Δλ of the resonance located at λ 0 , and to the cavity photon lifetime τ through the equation: Q= ! 0 !! ="# (2.19) where ω is the optical frequency (ϖ = 2πc/λ). The quality factor Q is often used to characterize the resonators because it is an important parameter, as it strongly determines the potential for applying to a cavity for many different areas. For examples, it plays an important role in determining the resonant bandwidth and loss for a passive WGM cavity filter or in the determination of thresholds for active processes, such as the lasing and nonlinear optical wave generation. Cavity loss is commonly expressed in terms of quality factor Q and there are several loss mechanisms in optical WGM resonators. The overall cavity quality factor Q is given by: 23 1 1 1 1 1 − − − − − + + + = w bulk ss rad Q Q Q Q Q (2.20) where rad Q is due to a purely radial loss for an ideal dielectric WGM resonator, the scattering losses from the residual surface in homogeneities ss Q , absorption losses due to contamination on the surface of the resonator w Q , and the bulk absorption in the fused silica bulk Q . The intrinsic material losses are known very accurately, since they arise from absorption in the material at the wavelength of concern [28]. Nevertheless, uncertainty is associated with the losses due to surface scattering and absorption, in which absorbed material on the surface of the resonator is contaminated, of which water is likely the principal culprit. The contribution to the quality factor for purely radiate effects, rad Q , can be derived by following the arguments presented in this Article [29]. These losses are due to the leakage of light from the resonator due to its finite dielectric constant and radius of curvature. The results can then be compared to numerical results obtained by l T / b rad e ) n ( n ) . l ( . Q 2 2 1 2 2 1 1 5 0 5 0 − + = − (2.21) where ( )( ) l l l tanh . l T η − η + = 5 0 ! l =arccosh n 1! 1 l+0.5 t q 0 "+ l 1!2b l 2 !1 " # $ % & ' ( ) * + , - !1 . / 0 1 0 2 3 0 4 0 ( ) [ ] 3 1 5 0 5 0 / . l . + = ξ 24 where b = 0 for TE modes and b = 1 for TM modes, n is index of refraction of resonator’s material, t q 0 is the q th zero of the airy function, q is corresponds to the radial mode number, in our case, we are interested in q = 1, and l is angular mode number. Let’s consider at wavelength λ = 1550 nm, the resonator’s radius around of 20 µm then 25 10 4 3 × ≈ . Q rad , thus with our resonator’s radius in the mm range, the net quality factor would be dominated by other loss mechanisms. Since it depends only on the absorption of the material at the wavelength of concern [30], the quality factor due to bulk absorption, bulk Q , in fused silica is actually known very well. αλ π = n Q bulk 2 (2.22) where n is the refractive index, and α is the absorption coefficient of the material. For example, a fused silica has a minimum, in its absorption coefficient, of 1 5 10 5 1 − − ≈ α m x . at 1550nm, which yields 11 10 9 3 × ≈ . Q bulk . Moreover, greater uncertainty is associated with the losses due to surface scattering and absorption due to surface roughness or the presence of an absorbing impurity on the surface of the disk. For the any given size of surface roughness and its correlated length, the surface scattering quality factor ss Q has to take into account not only the direct scattering light out of the disk but also the scattering into other modes with a high rate of leakage. We may apply the expression to describe the quality limit due to surface scattering [31]: 25 ( ) ( ) ( ) 2 2 2 1 2 7 2 5 3 2 1 4 2 3 B D Q / / / ss σ λ − ε π + ε ε = (2.23) where D is disk diameter, σ is the root-mean-square of the surface roughness, and B is the surface correlation length. We may estimate by using the numerical values as σ = 1nm, B = 5nm [33], then 13 10 ≈ ss Q . The quality factor due to the water adsorbed on the surface, Q w , is given by [32]: ) ( D n Q w / / w λ β δλ π ≈ 2 1 2 1 3 8 (2.24) where δ ∼ 0.2nm is an estimated thickness for the water layer, and 1 1183 − = λ β m ) ( is the absorption coefficient of water at wavelength 1550nm [33]. Because of the absorption coefficient of water at wavelength 1550 nm is so large, the Quality factor, due to the water absorbed and other contaminations on the surface, is dominant. It limits the intrinsic Q of resonator to the range of 10 8 . Thus, we can relax in the restriction for fabrication, which relates to the quality factor due to surface scattering. The root-mean- square of the surface roughness and the surface correlation length can be 10 nm and 50 nm, and results in 9 10 ≈ ss Q . 26 Figure 2.8: Scattering losses from residual surface in homogeneities ss Q , and the absorption losses due to a contamination on the surface of the resonator w Q as a function of resonator’s diameter. 2.4.2 The Free Spectral Range The free spectral range (FSR) of a cavity represents the frequency (wavelength) spacing between successive longitudinal modes. This definition, commonly used for Fabry-Perot (FP) cavities, is somewhat ambiguous for the cavities in this work, as the mode spectrum is highly complex. However, by using the most direct analogy to a FP, a microresonator can be considered a FP cavity wrapped around onto itself, such that the periphery of the resonator corresponds to the FP mirror spacing. The transmission response of the resonator is a periodic array of notches which occurring at the resonance Q_ss D ( ) 3 ε ⋅ ε 2 + ( ) 2 ⋅ 4π ( ) 3 ε 1 − ( ) 5 2 λ 7 2 D 1 2 σ B ⋅ ( ) 2 ⋅ := Q_w D ( ) π 8n 3 D 1 2 δ λ 1 2 ⋅ β_water ⋅ := 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 1 . 10 7 1 . 10 8 1 . 10 9 1 . 10 10 1 . 10 11 1 . 10 12 1 . 10 13 1 . 10 14 The resonator's diameter (m) The Quality Factors Q_ss D ( ) Q_w D ( ) D 27 wavelength. The separation between successive resonances is defined as free spectral range (FSR) and can be calculated by subtract two adjacent resonance mode number: m ) ( Rn m ) ( Rn m eff m eff m m λ π λ π λ λ 2 1 2 1 1 − + = − + + (2.25) The group index N usually is used because over larger range of wavelength, the propagation constant is not strictly constant due to the waveguide dispersion, thus λ λ β d dn n dk dn k n k N eff eff eff eff − = + = ∂ ∂ ≡ (2.26) then N ) ( n ) ( n m eff m eff ≈ ≈ + λ λ 1 and 2 1 λ λ λ ≈ × + m m , the FSR will be: eff Rn FSR π λ 2 2 ≈ (2.27) 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 5 . 10 11 1 . 10 10 1.5 . 10 10 2 . 10 10 2.5 . 10 10 3 . 10 10 Resonator's Radius (m) Free Spectrum Range (m) Figure 2.9: Free Spectrum Range as a function of Fused Silica resonator’s radius at wavelength 1550 nm 28 2.4.3 Finesse Finesse (F) is the ratio of the cavity mode spacing to the cavity bandwidth. This definition factors in both the cavity loss (Q) and cavity mode spacing (FSR) to obtain a dimensionless single parameter which characterizes the ability to resolve the cavity resonance structure, as given by F= FSR FWHM = FSR!Q ! (2.28) Knowledge of cavity Finesse is useful for a wide range of subjects, as it determines several properties, for example, the amount of spectral noise/power a resonance filter passes, and the amount of energy amplification in a resonance system. Cavity Finesses in ultra high Q whispering gallery mode microresonators can exceed 10 6 [30, 32]. 29 2.5 Chapter References [1] J. W. S. Rayleigh, The Theory of Sound, Volume Two. New York: Dover, 1945. [2] L. Rayleigh, “Further applications of Bessel’s functions of high order to the Whispering Gallery and applied problems,” Philosophical Magazine, vol. 27, pp. 100-109, 1914. [3] L. Rayleigh, “The problem of the Whispering Gallery,” Philosophical Magazine, vol. 20, pp. 1001-1004, 1910. [4] V. B. Braginsky, M. L. Gorodetsky, and V. S. Ilchenko, “Quality-factor and nonlinear properties of optical whispering-gallery modes,” Physics Letters A, vol. 137, pp. 393-396, 1989. [5] A. A. Savchenkov, A. B. Matsko, V. S. IIchenko, and L. Maleki, “Optical resonators with ten million finesse,” Optics Express, vol. 15, pp. 6768-6773, 2007. [6] M. L. M. Balistreri, D. J. W. Klunder, F. C. Blom, A. Driessen, H. W. J. M. Hoekstra, J. P. Korterik, L. Kuipers, and N. F. Hulst, “Visualizing the whispering gallery modes in a cylindrical optical microcavity,” Optics Letters, vol. 24, pp.1829-1831, 1999. [7] D. K. Armani, T. J. Kippenberg, S. M. Spillane, and K. J. Vahala, “Ultra-high-Q toroid microcavity on a chip,” Nature, vol. 421, pp. 925-928, 2003. [8] B. E. Little, J. S. Foresi, G. Steinmeyer, E. R. Thoen, S. T. Chu, H. A. Haus, E. P. Ippen, L. C. Kimerling, and W. Greene, “Ultra-compact Si–SiO 2 microring resonator optical channel dropping filters,” IEEE Photonics Technology Letters, vol. 10, pp. 549-551, 1998. [9] P. Rabiei, W. H. Steier, C. Zhang, and L. R. Dalton, “Polymer micro-ring filters and modulators,” Journal of Lightwave Technology, Vol. 20, pp. 1968-1975, 2002. [10] D. W. Vernooy, A. Furusawa, N. P Georgiades, V. S Ilchenko, and H. J. Kimble, “Cavity QED with high-Q whispering gallery modes,” Physical Review A, vol. 57, pp. R2292-R2296, 1998. [11] V. Sandoghdar, F. Treussart, J. Hare, V. Lefevre-Seguin, J. M. Raimond, and S. Haroche, “Very low threshold whispering-gallery-mode microsphere laser,” Physical Review A, vol. 54, pp. R1777-R1780, 1996. [12] M. Cai, O. Painter, K. J. Vahala, and P. C. Sercel, “Fiber-coupled microsphere laser,” Optics Letters, vol. 25, pp.1430-1432, 2000. 30 [13] S. Arnold, M. Khoshsima, I. Teraoka, S. Holler and F. Vollmer, “Shift of whispering-gallery modes in microspheres by protein adsorption,” Optics Letters, vol. 28, pp. 272-274, 2003. [14] H. Quan and Z. Guo, “Simulation of single transparent molecule interaction with an optical microcavity,” Nanotechnology, vol. 18, 2007. [15] C. G. B Garrett, W. Kaiser, and W. L. Bond, “Stimulated emission into optical whispering gallery modes of spheres,” Physical Review, vol. 124, pp. 1807-1809, 1961. [16] J. C. Knight, H. S. T. Driver, R. J. Hutcheon, and G. N. Robertson, “Core resonance capillary fiber whispering gallery mode laser,” Optics Letters, vol. 17, pp. 1280-1282, 1992. [17] V. S. Ilchenko, A. A. Savchenkov, A. B. Matsko, and L. Maleki, “Nonlinear optics and crystalline whispering gallery mode cavities,” Physical Review Letters, vol. 92, 2004. [18] A. A. Savchenkov, V. S. Ilchenko, A. B. Matsko, and L. Maleki, “KiloHertz optical resonances in dielectric crystal cavities,” Physical Review A, vol. 70, 2004. [19] I. Teraoka, S. Arnold, and F. Vollmer, “Perturbation approach to resonance shifts of whispering-gallery modes in a dielectric microsphere as a probe of a surrounding medium,” Journal of the Optical Society of America B, vol. 20, pp. 1937-1946, 2003. [20] N. M. Hanumegowda, C. J. Stica, B. C. Patel, I. White and X. Fan, “Refractometric sensors based on microsphere resonators,” Applied Physics Letters, vol. 87, 2005. [21] A. M. Armani and K. J. Vahala, “Heavy water detection using ultra-high-Q microcavities,” Optics Letters, vol. 31, pp. 1896-1898, 2006. [22] A. A. Savchenkov, A. B. Matsko, M. Mohageg, and L. Maleki, “Ringdown spectroscopy of stimulated Raman scattering in a whispering gallery mode resonator,” Optics Letters, vol. 32, pp. 497-499, 2007. [23] F. Vollmer, and S. Arnold, “Whispering-gallery-mode biosensing: label-free detection down to single molecules,” Nature Methods, vol. 5, pp. 591-596, 2008. [24] M. Sumetsky, R. S. Windeler, Y. Dulashko, and X. Fan, “Optical liquid ring resonator sensor,” Optics Express, vol. 15, no. 22, Oct. 2007. [25] A. M. Armani, R. P. Kulkarni, S. E. Fraser, R. C. Flagan, and K. J. Vahala, “Label- Free, Single-Molecule Detection with Optical Microcavities,” Science, vol. 317, pp. 783-787, 2007. 31 [26] A. Yariv, Optical Electronics, 4 th ed. Holt McDougal, 1991. [27] M. Oxborrow, “Traceable 2D finite-element simulation of the whispering-gallery modes of axisymmetric electromagnetic resonators,” IEEE Transactions on Microwave theory and Techniques, vol. 55, no. 6, pp. 1209-1218, Jun. 2007. [28] K. Chang, Handbook of Microwave and Optical Components. New York: Wiley, 2001. [29] V. V. Datsyuk, “Some characteristics of resonant electromagnetic modes in a dielectric sphere,” Applied Physics B: Lasers and Optics, vol. 54, no. 2, pp. 184-187, 1992. [30] M. L. Gorodetsky, A. A. Savchenkov, and V. S. Ilchenko, “Ultimate Q of optical microsphere resonators,” Optics Letters, vol. 21, no. 7, pp. 453-455, 1996. [31] K. K. Lee, D. R. Lim, L. C. Kimerling, J. Shin, and F. Cerrina, “Fabrication of ultralow loss Si/SiO 2 waveguides by roughness reduction,” Optics Letters, vol. 26, no. 23, pp. 1888-1890, 2001. [32] D. W. Vernooy, V. S. Ilchenko, H. Mabuchi, E. W. Streed, and H. J. Kimble, “High- Q measurements of fused-silica microspheres in the near infrared,” Optics Letters, vol. 23, no. 4, pp. 247-249, 1998. [33] L. Matthies, P. Bellutta, M. McHenry,“Detecting water hazards for autonomous off- road navigation,” Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA, USA 91109, 2003. 32 Chapter 3 Polymer Optical Waveguide Vertically Coupled to High Q Whispering Gallery Resonators 3.1 Introduction Optical WGM resonators have numerous applications in sensing [1, 2], nonlinear optics [3, 4], and photonics [5, 6]. Fused silica sphere and disc micro-resonators have been reported with Q’s exceeding [3, 4, 7]. Tapered fibers, side polished fibers and frustrated total internal reflection prisms are typically used to couple to the whispering gallery modes in these resonators. On the other hand, ring waveguide and disc micro- resonators have been integrated with coupling bus waveguides in several materials systems [8-10]. The highest reported Q’s in these resonators are ~10 5 , usually limited by scattering from sidewalls. We proposed and experimentally demonstrated to incorporate the high Q fused silica discs with the polymer integrated circuit technology. This technology will widely open the ability of development integrated photonics devices, sensing array implementation for parallel detection and portable devices. By appropriate design of polymer waveguide-fused silica resonator coupling, a complete 100% transmission of energy from the polymer waveguide to high Q resonator is possible [11]. The main challenges are to design the structures of waveguide and resonator to achieve the efficient coupling and easy to implement into optical integration technology. Polymer and Fused Silica have very close refractive index, thus it help us to easily achieve small 33 phase mismatch between the waveguide mode and WGM mode in the resonator. It leads to transfer energy efficiently from waveguide to resonator disk and critical coupling is possible. 3.2 Theoretical Approach of Polymer Waveguides Coupled to WGM Resonators Figure 3.1: Optical polymer waveguide coupled to WGM resonator The diagram of polymer waveguide optically vertically coupled to high Q Fused Silica resonator disk is shown in figure 3.1. We can follow closely the reference [12] to perform the analysis of our model. It obeys the coupled mode equation with the solution given by ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ⎥ ⎥ ⎦ ⎤ − ⎢ ⎣ ⎡ = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ∗ ∗ 2 1 2 1 A A t t B B κ κ (3.1) 34 Where ) i exp( t t t φ ⋅ = is the field amplitude that passes the coupler, t φ is the phase mismatch introduces by coupling between the waveguide and resonator and κ is the field amplitude that couples into or out of resonator disk. The optical field 2 B that couples into the resonator has a phase accumulation and amplitude decrease that result in an input to the coupler 2 A , we have 2 2 B ) i exp( A ⋅ ⋅ = θ α (3.2) α and θ are respectively the loss and the phase shift after one circulation. L β θ = with β is propagation constant which depends on the effective refractive index of resonator as λ π β eff n 2 = and the length L = 2πR. The normalized transmitted power at the output is ) cos( t t ) cos( t t A B T t t φ θ α α φ θ α α + − + + − + = = 2 1 2 2 2 2 2 2 1 1 (3.3) In principle, if the coupling length is short enough then t φ θ >> , first approximation one can disregard the affect of t φ and the perturbation mainly by the wave propagation along the cavity. The amount of power coupled from the waveguide into the resonator is proportional to ) exp( 2 β Δ γ ⋅ − , where γ is a constant decided by the propagation constant of light in the waveguide and the diameter of the resonator (coupling length). β Δ is difference in propagation constant between waveguide and 35 resonator which effect the coupling efficiency. When 0 = β Δ we have the phase matching condition and the coupling efficiency is maximized. When the mode is on resonance, the coupling can be considered into three regimes: (1) Under-coupled regime: when the waveguide is far from resonator, the coupling is very weak. When the waveguide moves close to resonator, the overlap of the waveguide mode and WGMs increase. However, α > t the transmission T decreases continuously from unity and approach to zero. (2) Critical coupled regime: when the waveguide get closer to resonator such that α = t , the normalized transmission T is zero, which means that all the input power is coupled to the resonator. (3) Over-coupled regime: When the gap between waveguide and resonator is further decreased, the overlap of modes become larger and the normalized transmission T become greater than 0 again, it is called over-coupled regime. By appropriate design of polymer waveguide-fused silica resonator coupling, a complete 100% transmission of energy from the polymer waveguide to high Q resonator is possible [11]. The main challenges are to design the structures of waveguide and resonator to achieve the efficient coupling and easy to implement into optical integration technology. Polymer and Fused Silica have very close refractive index, thus it help us to easily achieve small phase mismatch between the waveguide mode and WGM mode in the resonator. It leads to transfer energy efficiently from waveguide to resonator disk and critical coupling is possible. 36 1.550352 . 10 6 1.550353 . 10 6 1.550354 . 10 6 1.550355 . 10 6 1.550356 . 10 6 1.550357 . 10 6 0 0.2 0.4 0.6 0.8 1 Wavelength (m) Normalized Transmission Output Figure 3.2: Normalized Transmission Output in different coupling regimes Under-coupled regime (green), Critical coupling (red) and Over-coupled regime (blue) 3.3 Design and Fabrication Optical Polymer Waveguides and High Q WGM Fused Silica Resonator Disks In this approach, the fused silica (Suprasil F300) disc, radius 1.6 mm, thickness 35 m µ , is fabricated with the sides optically flat and the edge optically polished. On a separate Si substrate a group of straight polymer waveguides are fabricated. The waveguides are designed to have the same effective index of refraction as the mode in the fused silica disc at λ = 1550 nm. One of the waveguides is vertically coupled to the mode of the disc by careful alignment near the perimeter of the disc, and by adjusting the vertical air gap between the disc and the waveguide. 37 From the one square inch, 500um thick Fused Silica material, we use our own fabricated spin saw to cut a cylinder disc of diameter about 5mm. Next, we use Precision Polishing TechPrep System Model #15-2000 (figure 3.3) from Allied High Tech Product, Inc with diamond lapping films (range from 30 um to 0.1 um) to form the disc thickness about 50um with both top and bottom surfaces optically polished, however, the edge is still very rough. We then use thermal wax to glue the disc on top of brass holder as in figure 3.4. With this configuration we are able to mechanically spin and optically polish the edge of disc by latch in our machine shop. Figure 3.3: Precision Polishing TechPrep System Model #15-2000 from Allied High Tech Product, Inc 38 Figure 3.4: High Q WGM Fused Silica resonators 1 10 3 − × 2 10 3 − × 3 10 3 − × 4 10 3 − × 5 10 3 − × 1.434 1.436 1.438 1.44 1.442 Resonator's Radius (m) Effecitve Refractive Index Figure 3.5: The plot is the effective refractive index of WG fundamental modes TE (red) and TM (blue) as a function of the resonator’s radius, with resonator’s material is Fused Silica with refractive index n = 1.444 at wavelength λ = 1550 nm and surround medium is air 39 Efficient coupling from the polymer waveguide to the fused silica resonator requires a velocity match between the waveguide mode and the whispering gallery mode of the disc. The effective index of the lowest order radial mode of the disc was calculated using the equation (2) in [4] and the results are shown in figure 3.5 for a wavelength of 1550 nm. The estimated effective index of a 1.6 mm radius disc is 1.4384. We invoke the numerical tool to predict the waveguide optical properties which called Effective index method. The details of illustration this method can be referenced in [13]. The general concept can be summarized as following steps: (1) The normalized thickness of channel f V and lateral l V region are defined as 2 3 2 1 n n kt V l , f l , f − = (3.4) (2) Define asymmetry parameter 2 2 2 1 2 3 2 2 n n n n a − − = (3.5) (3) Numerical solution for l , f b are found using the equation (for TE case) l , f l , f l , f l , f l , f l , f b a b arctan b b arctan b V − + + − + = − ⋅ 1 1 1 νπ (3.6) (4) The equivalent effective indices l , f N of three regions given by: (N f,l ) 2 =n 3 2 +b f,l (n 1 2 !n 3 2 ) (3.7) (5) Finding the normalized thickness: 2 2 l f eq N N kw V − = (3.8) 40 (6) Solving for the normalized guide index eq eq eq eq eq eq b a b arctan b b arctan b V − + + − + = − 1 1 1 νπ (3.9) (7) Finally, solve for the effective refractive index of the waveguide ) N N ( b N n l f eq l eff 2 2 2 − + = (3.10) Figure 3.6: The field intensity of polymer waveguide was calculated by Effective Index method The single mode polymer waveguides were fabricated on a Si substrate by spin coat a 7 m µ thickness lower cladding of ZPU13-R1 with a refractive index of 1.4053 @ 1550 nm. The core material is ZPU12-R1 with a refractive index of 1.4701 @ 1550 nm. ZPU13-R1 and ZPU12-R1 were supplied by ChemOptics in Korea. The upper cladding is air and a ridge waveguide is used for lateral confinement. 5 4 3 2 1 0 1 2 0 0.2 0.4 0.6 0.8 1 Dimension (um) Normalized Intensity, a.u Lower Cladding Core Air 5 4 3 2 1 0 1 2 0 0.2 0.4 0.6 0.8 1 Dimension (um) Normalized Intensity, a.u Lower Cladding Core Air 41 Figure 3.7: The fabrication procedure and dimension of polymer waveguide 42 The fabrication procedure and dimensions of polymer waveguide are as in figure 3.7. The Effective index method was used to design the waveguide with an effective index close to that of the disc but with a width corresponding to our existing masks. With a waveguide width of 3.5 m µ , a ridge height of 0.5 m µ and a core width (outside the ridge) of 1.4 m µ , the calculated effective index is 1.4376. Over the estimated effective coupling length between the waveguide and the disc of ~150 m µ , this index mismatch is negligible. We also used the Multiphysics Modeling and Simulation Software from COMSOL with Finite Element Method (FEM) to model our polymer waveguide. Figure 3.8 shown the field distribution of polymer waveguide and also calculate the effective mode index of 1.436724 at wavelength of 1550nm which agreed with effective refractive index 1.4376 from analytical calculation Figure 3.8: Field Intensity distribution of polymer waveguide was simulated by Multiphysics Modeling and Simulation Software from COMSOL with Finite Element Method (FEM). 43 Figure 3.9: Optical polymer waveguide was fiber butt coupled by single mode fiber Upper: SMF was butt coupled into one of two polymer waveguides in the same substrate Lower: Red laser 634nm was launched into polymer waveguide by fiber butt coupling. 44 3.4 Experiment and Results The disc was glued by wax on top of the brass holder post and the polymer waveguides were mounted on a micro-manipulator for both horizontal and vertical adjustment. The waveguides were placed under the disc. One of the waveguides as shown in figure 3.9 was positioned near the perimeter of the disc and fine tune the lateral position by the help of a piezo-electric actuator with a 3 VDC per 100 nm movement for maximum coupling. A tunable 1550 nm laser was fiber butt coupled into the TM mode of the correct waveguide and the waveguide output was fiber coupled to a photo-detector. The vertical coupling gap between the waveguide and the disc was adjusted with the micro-manipulator and fine tuned by a piezo-electric actuator. Figure 3.10: Overview illustration of Polymer waveguide vertically coupled to high Q WGM resonators. Laser PhotoDetector Polymer Waveguide WGM resonator 45 Figure 3.11: Experimental demonstration of optical Polymer waveguide vertically coupled to high Q WGM Fused Silica resonator disk. 46 Figure 3.12 (Upper) shows the output power vs. detuning frequency and shows coupling to different longitudinal modes and radial modes of the disc. The frequency spacing between the sets of the same radial modes is consistent with the calculated free spectral range (FSR) of 20.65 GHz, where FSR = c/2πn eff R. Figure 3.12 (Lower) shows one mode on an expanded frequency scale. The measured loaded Q is € 1.2×10 7 and the coupling is very near critical coupling. If we assume critical coupling, the unloaded Q is € 2.4×10 7 . Based on this, the total propagation loss (the sum of scattering loss due to roughness, absorption loss of material, and radiation loss due to bending) is about 3 10 5 . 8 − × 1 cm − that leads to the roundtrip loss 0.9% in the disc. If surface scattering is dominant, the surface roughness is in the nanometer range [14]. The power coupled to one of the modes was measured as a function of the air gap when the wavelength was tuned to be on resonance, and is shown in figure 3.13. A gap between the optical coupler and the WGM resonator with a uniform index is generally needed to achieve a proper optical coupling. This gap is used to properly "unload" the WG mode. The Q factor of a WG mode is determined by properties of the dielectric material of the WGM resonator, the shape of the resonator, the external conditions, and strength of the coupling through the coupler (e.g. polymer waveguide). The highest Q factor may be achieved when all the parameters are properly balanced to achieve a critical coupling condition. In WGM resonators with uniform indices, if the coupler such as an optical waveguide touches the exterior surface of the resonator, the coupling is strong and this loading can render the Q factor to be small. Hence, the gap between the 47 Figure 3.12: Optical transmission spectrum for the polymer waveguide vertically coupled to Fused Silica resonator disk system at 1550nm. (Upper) Output power versus laser detuning frequency. (Lower) Expanded frequency scale is showing a Loaded Q of 1.2 x 10 7 . d = 3.2 mm, FSR = 20.65GHz 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 5 10 15 20 25 30 35 Laser Frequency Tuning (Ghz) Normalized Output Intensity, a.u. Q = 1.2*10^7, d = 3.2mm 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 50 100 150 200 250 300 350 400 450 500 Laser Frequency Tuning (MHz) Normalized Output Intensity, a.u. 16.52 MHz 48 Figure 3.13: Transmission factor versus the gap between the optical polymer waveguide and WGM resonator disk. surface and the coupler is used to reduce the coupling and to increase the Q factor. In general, this gap is very small, e.g., less than one wavelength of the light to be coupled into a WG mode. Precise positioning devices such as piezo-electric elements may be used to control and maintain this gap at a proper value. In our experiment demonstration, we were able to keep the bottom of the disc and the top of the waveguides parallel and to estimate the point when the top of the polymer waveguide touched the disc by observing the contact fringes through a microscope. Critical coupling is obtained when the gap is ~0. 4 m µ . Using coupled mode theory [15] and our calculated mode patterns in the disc and the waveguides, critical coupling occurs at 0.5 m µ gap, in reasonable agreement with the experiment. 0 10 20 30 40 50 60 70 80 90 100 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Gap between polymer waveguide and resonator disc (um) Transmission Factor (%) 49 3.5 The Integrated Devices Demonstration Integrated Optics is an important technology that will be a key in many advanced photonics and communication applications. For instance, in communication optical fiber link, the need of progress in integrated optics to accomplish the compactness, reliability and cost reducing. In photonics sensor, the compactness, simplicity and potential for mass production of integrate optics will be great benefits. The semiconductor integrated optics wavelength selective is commercial device, however, the bandwidth is still broad. We first time demonstrated a novel polymer optical waveguide integrated to high Q whispering gallery fused silica resonator which function as an integrated optics narrow bandwidth wavelength selective devices. The device is basic for integration to form narrow band receiver for wavelength division multiplexed optical communication systems. The device may be suitable for monolithic integration with a photodetector as the drop port to form a wavelength demultiplexing receiver. Owing to the flexibility of fabricating polymer waveguide and whispering gallery resonator we can freely chose the center wavelength and designed bandwidth. In the process of integration of the silica disc with the polymer waveguide, we used 6-D micro- manipulator alignment system with piezo-electric actuators to align and achieve the desired coupling, then used curable UV epoxy to attach both aluminum substrates to form an integrated optics narrow bandwidth wavelength selective device. By the use of 6-D micro-manipulator alignment system with piezo-electric actuators we were able to keep the bottom of the disc and top of the waveguides parallel and control the air gap between the resonator and polymer waveguide so as the power coupling more precise than our 50 first experiment. We achieved zero transmission factor at critical coupling with unloaded Q = 2.4 x 10 7 as in the figure 3.12. After we used the curable UV epoxy to bond both substrates, we can separate the integrated wavelength selective chip from the micro- manipulator alignment system. All device then package in the rectangular box with input and output in fiber pigtail form. Figure 3.15 shows the output power vs. detuning frequency. The optical transmission spectrum showed: transmission factor now is about 12% and loaded Q = 4 x 10 6 . We can recognize that the coupling in over-coupled regime because the gap between the polymer waveguide and resonator disc is smaller due to tiny shrinkage of curable UV epoxy. Figure 3.14: Integration of polymer waveguide vertically coupled to high Q WGM resonator in a practical packet. 51 Figure 3.15: Integration of polymer waveguide vertically coupled to high Q WGM resonator: Normalized Transmission Output Intensity versus Laser detuning frequency. We have the first experimental demonstration of direct coupling between a high Q resonators and a polymer waveguide. This configuration allows the use of very high Q micro-resonators in integrated optics circuit. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 50 100 150 200 250 300 350 400 450 500 Normalized Output Intensity, a.u. Laser Frequency Detuning (MHz) Intergrated Polymer Waveguide with WGM resonator Loaded Q = 4.7*10 6 , Coupling efficiency of 88% 41.5 MHz 52 3.6 Chapter References [1] F. Vollmer, D. Braun, A. Libchaber, M. Khoshsima, I. Teraoka, and S. Arnold, “Protein detection by optical shift of a resonant microcavity,” Applied Physics Letters, vol. 80, pp. 4057-4059, 2002. [2] B. Bhola, H. C. Song, H. Tazawa, and W. H. Steier, “Polymer microresonator strain sensors,” IEEE Photonics Technology Letters, vol. 17, no. 4, pp. 867-869, Apr. 2005. [3] V. B. Braginsky, M. L. Gorodetsky, and V. S. Ilchenko, “Quality-factor and nonlinear properties of optical whispering-gallery modes,” Physics Letters A, vol. 137, no. 7-8, pp. 393-397, May 1989. [4] V. S. Ilchenko, A. A. Savchenkov, A. B. Matsko, and L. Maleki, “Non-linear optics and crytalline whispering gallery mode cavities,” Physical Review Letters, vol. 92, no. 4, Jan. 2004. [5] S. J. Choi, Z. Peng, Q. Yang, and P. D. Dapkus, “An eight channel demultiplexing switch array using vertically coupled active microdisk resonators,” IEEE Photonics Technology Letters, vol. 16, no. 11, pp. 2517-2519, Nov. 2004. [6] V. Sandoghdar et al., “Very low threshold whispering-gallery-mode microsphere laser,” Physics Review. A, vol. 54, pp. R1777-R1780, 1996. [7] D. K. Armani, T. J. Kippenberg, S. M. Spillane, and K. J. Vahala, “Ultra high-Q toroid microcavity on a chip,” Nature, Vol. 421, pp. 925-928, Feb. 2003. [8] B. E. Little, S. T. Chu, W. Pan, D. Ripin, T. Kaneko, Y. Kokubun, and E. Ippen, “Vertically coupled glass microring resonator channel dropping filters,” IEEE Photonics Technology Letters, vol. 11, no. 2, pp. 215-217, Feb. 1999. [9] P. Rabiei, W. H. Steier, C. Zhang, and L. R. Dalton, “Polymer micro ring filters and modulators,” Journal of Lightwave Technol.ogy, vol. 20, no. 11, pp. 1968-1975, Nov. 2002. [10] D. V. Tishinin, P. D. Dapkus, A. E. Bond, I. Kim, C. K. Lin, and J. O’Brien, “Vertical resonant couplers with precision coupling efficiency control fabricated by wafer bonding,” IEEE Photonics Technology Letters, vol. 11, no. 8, pp. 1003-1005, Aug. 1999. [11] A. Yariv, “Universal relations for coupling optical power between microresonator and dielectric waveguides,” Electronics Letters, vol. 36, no. 4, 2000. [12] A. Yariv, “Critical coupling and its control in optical waveguide-ring resonator systems,” IEEE Photonics Technology Letters, vol. 14, no. 4, Apr. 2002. 53 [13] T. Tamir, Guided-Wave Optoelectronics (Springer Series in Electronics and Photonics). Springer-Verlag, 1990. [14] M. Borselli, T. J. Johnson, and O. Painter, “Beyond the Rayleigh scattering limit in high-Q silicon microdisk: Theory and experiment,” Optics Express, vol. 13, no. 5, pp. 1515-1530, Mar. 2005. [15] D. Marcuse, Theory of Dielectric Optical Waveguides. New York: Academic, 1974. 54 Chapter 4 Differential TE and TM Mode Measurement: A Method to Reduce The Effect of Thermal Drift in Optical Resonant Sensors 4.1 Introduction Recently reported progress in the fabrication of ultra high Q resonators has motivated the development of compact, integrated, and very high sensitivity sensors for physical, chemical, and biological detection [2-6]. In these devices, the evanescent fields of WGMs in resonators outside dielectrics strongly interact with the surrounding medium. As a result, the WGM resonance frequency is extremely sensitive to changes in refractive index of the surrounding medium due to the presence of chemical substances, the adsorption of a molecular layer, and even the attachment of a single molecule [4-6, 23-25]. One of the common detection schemes monitors the shift of the resonance frequency that is proportional to the quantitative medium change in the surrounding [3, 4]. However, temperature fluctuations also cause shifts of the resonance frequencies and thus limit the practical detection sensitivity. The large thermal drift makes this method difficult to use outside the laboratory and often requires active temperature stabilization. The purpose of this work will be to investigate and experimentally demonstrate a method for significantly reducing the effect of thermal drift in WGM resonator based sensors without the need of temperature stabilization. The technique exploits the difference 55 between the TE and TM mode resonant frequencies. These two modes have nearly identical changes with respect to temperature yet significant difference in sensitivity to the changes in the material surrounding the resonator. Therefore, the method based on the difference between TE and TM mode frequencies leads to a practically achievable low detection limit in the presence of temperature changes. Consequently, the United States Patent 8,111,402 was granted for our novel invention. 4.2 Theoretical Principle The WGM wavelength resonance is related to the effective refractive index as , where R is the resonator’s radius, n eff is the effective refractive index of the WGM resonator, λ is the wavelength of light, and l is an integer that describes the WGM angular mode number. For a given WGM mode, the temperature induced frequency shift can be obtained as: (4.1) The first term in the above equation accounts for the thermal expansion effect with is the linear thermal expansion coefficient. The second term in the equation describes the effect of the temperature fluctuation in the effective refractive index of resonator, n eff . In general, n eff depends on polarization, refractive index of the material, refractive index of surrounding medium, and the geometry of the resonator. The change of the effective refractive index due to temperature fluctuation can be described as: ! l = 2"Rn eff # ! "f f TE,TM = #R #T $ 1 R "T + #n eff #T $ 1 n eff "T ! " = 1 R #R #T 56 (4.2) where and are the thermo optics coefficients for the surrounding medium and the resonator’s material, n is the refractive index of the resonator’s material, and n o is the refractive index of the surrounding medium, respectively. The last term is the change in the effective refractive index of the resonator due to thermal expansion [26]. Our study focuses on Fused Silica resonators. The thermal expansion coefficient of fused silica is = 5.5 x 10 -7 / o C. This is small compared to the thermo-optic coefficient which is = 1.28 x 10 -5 / o C [27] for fused silica, = -10 -4 / o C for water and = 10 -6 / o C for air. One can ignore the thermal expansion term in equations (4.1) and (4.2), since the method of using the differential frequency measurements should completely eliminate the effect of thermal expansion. The method is, therefore, useful even for large thermal expansion coefficient materials. For the low order radial modes in a large disc resonator which are typically used in these experiments, the resonant frequency, w n,q and the effective index can be approximated as shown in [28]: (4.3) (4.4) ! "n eff "T =# out "n eff "n o +# n "n eff "n +#R "n eff "R ! " out = #n o #T T n n ∂ ∂ = α α n α ! " out ! " air ! " #,q c $ 1 Rn o # m + % q m # 2 & ' ( ) * + 1/3 , P m 2 ,1 + 3% q 2 20m 2 # & ' ( ) * + 1/3 +O(# ,2/3 ) - . / 0 1 2 q , v vq eff R vc n ω ≈ 57 where ν = l + with l = 0, 1, 2…is an angular momentum number, c is a speed of light in vacuum, P = 1 for the case of TE mode or P = for the case of TM mode, m = , is the qth root of the Airy function. When the temperature changes by ΔT, the refractive index of resonator’s material changes to n + Δn = n + α n ΔT and surrounding refractive index changes to n o + α out ΔT. Combined with equation (4.3), we can approximate the expression for resonance frequency shift as a function of the temperature fluctuation for TE mode: + (4.5) and similarly for TM mode, we get + (4.6) 2 1 2 1 m o n n q α ! "# TE c Temp $ 1 R n n 2 %n o 2 ( ) 3/2 % 1 n 2 & +' q & 2 ( ) * + , - 1/3 + 3' q 2 20 2 & ( ) * + , - 1/3 . / 0 1 2 3 ( ) * * + , - - ' n "T ! 1 R n o n 2 "n o 2 ( ) 3 2 # $ % % & ' ( ( ) out *T ! "# TM c Temp $ 1 R n o 2 3n 2 %2n o 2 ( ) n 3 n 2 %n o 2 ( ) 3/2 % 1 n 2 & +' q & 2 ( ) * + , - 1/3 + 3' q 2 20 2 & ( ) * + , - 1/3 . / 0 1 2 3 ( ) * * + , - - ' n "T ! 1 R n o n 2 "n o 2 ( ) 3/2 2" n o 2 n 2 # $ % & ' ( ) * + + , - . . / out 0T 58 The difference gives: (4.7) Consider the ratio and its two factors: (1 – 2(n o 2 /n 2 )) in the numerator and dominant large factor ν in the denominator, the magnitude of the ratio is about 10 -3 to 10 -5 with a large disc resonator. It shows that the differential resonance shift is significantly smaller than individual TM or TM resonance shift under the effect of temperature. Moreover, for the particular case of a fused silica disc surrounded by air, n o 2 /n 2 ~ ½ and the effect of temperature is very small as shown in Figure 4.1 (upper plot). Figure 4.1 (lower plot) is the theoretical calculation of the resonance shifts due to temperature fluctuation in the case of a Fused Silica disc surrounding by water. We note that the resonance frequency shift of each mode TE or TM due to temperature fluctuation is 5 orders of magnitude larger than the change of the differential frequency of two resonance modes for the case of air on the outside, and approximately 3 orders of magnitude larger for the case of water on the outside. In both cases the effects of temperature are significantly reduced. ! "# TE $"# TM ( ) c Temp % 1 Rn 1$2 n o 2 n 2 n 2 $n o 2 & ' ( ( ( ) * + + + , n "T + 1 Rn 2 n o n 2 $n o 2 & ' ( ( ) * + + , out "T Temp TE TM TE ω Δ ω Δ ω Δ − 59 Figure 4.1: The plot illustrates resonance frequencies shift due to temperature fluctuation for individual TE or TM mode and the change of differential frequency of two resonance modes. The theoretical calculation is based on 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, for the case of angular mode l = 9331, radial mode q = 1. Upper: The surrounding medium is air (n 0 = 1). Lower: The surrounding medium is water (n 0 = 1.333). Resonance wavelength shifts of TM and TE modes and the difference due to temperature fluctuation in case of air surrounding medium n 0 = 1 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × 1 10 3 × Temperature Fluctuation (C) Resonance wavelength shift (pm) TM/TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to temperature fluctuation in case of air surrounding medium n 0 = 1 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × 1 10 3 × Temperature Fluctuation (C) Resonance wavelength shift (pm) TM/TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to temperature fluctuation in case of water surrounding medium n 0 = 1.333 0 2 4 6 8 10 1.43835 1.4384 1.43845 1.4385 n_eff_TE ΔT ( ) ΔT 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × 1 10 3 × Temperature Fluctuation (C) Resonance wavelength shift (pm) TM/TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to temperature fluctuation in case of water surrounding medium n 0 = 1.333 0 2 4 6 8 10 1.43835 1.4384 1.43845 1.4385 n_eff_TE ΔT ( ) ΔT 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × 1 10 3 × Temperature Fluctuation (C) Resonance wavelength shift (pm) TM/TE Difference 60 To estimate the sensor’s detection limit, we assume a resulting ∆n o due to the uniform change in surrounding refractive index. We can calculate the shift of the resonant frequency of the individual modes by using equation (4.3). The approximation for TE modes is , (4.8) and similarly for TM mode is . (4.9) Thus the difference is: (4.10) The shift of the TE and TM modes and the difference between them due to the change in outside refractive index are calculated for the case of air surrounding medium shown in figure 4.3 (upper plot). With a typical temperature fluctuation of 1ºC during a measurement that leads to the resonance wavelength peak of TE or TM mode fluctuates about 13 pm as in figure 4.1, then the minimum detectable Δn o using the conventional method of tracking the change in resonance peaks of the individual TE or TM mode is 4 x 10 -2 RIU (Refractive Index Unit) of surrounding medium’s refractive index. On the other hand, if using the difference measurement method, one can detect a change as small as 2 x 10 -6 RIU. The detection limit is significantly improved by more than four orders of magnitude. A similar calculation for the case where the surrounding medium is water, ! "# TE c n_out $ 1 R n o "n o n 2 %n o 2 ( ) 3 2 ! "# TM c n_out $ 1 R n o "n o n 2 %n o 2 ( ) 3/2 2% n o 2 n 2 & ' ( ) * + ! "# TE $"# TM ( ) c n_out % 1 Rn 2 n o n 2 $n o 2 & ' ( ( ) * + + "n o 61 demonstrates that this is almost a two orders of magnitude increase in detection limit, as shown in the figure 4.2. If it is only 0.01 o C temperature fluctuation and air surrounding medium, our proposed method of using the differential frequency between TE and TM mode can detect the change of surrounding refractive index in the range of 10 -8 instead of about 3 x 10 -4 if we use the frequency shift of each individual TE or TM mode as shown in figure 4.3 (lower plot). Figure 4.2: The plot illustrates resonance frequencies shift due to the change in refractive index of surrounding medium for individual TE or TM mode and the change of differential frequency of two resonance modes. The theoretical calculation is based on 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, for the case of angular mode l = 9331, radial mode q = 1. The surrounding medium is water (n 0 = 1.333). Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of water surrounding medium n 0 = 1.333 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of water surrounding medium n 0 = 1.333 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference 62 Figure 4.3: The plot illustrates resonance frequencies shift due to the change in refractive index of surrounding medium for individual TE or TM mode and the change of differential frequency of two resonance modes. The theoretical calculation is based on 1.6 mm diameter fused silica resonator with n = 1.444 at 1550 nm, for the case of angular mode l = 9331, radial mode q = 1. The surrounding medium is air (n 0 = 1), and the lower plot is the expanding calculation of the change of outside refractive index to 10 -8 RIU. Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of air surrounding medium n 0 = 1 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of air surrounding medium n 0 = 1 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × 1 10 1 × 1 10 2 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of air surrounding medium n 0 = 1 1 10 8 − × 1 10 7 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of air surrounding medium n 0 = 1 1 10 8 − × 1 10 7 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × The change in outside refractive index Resonance wavelength shift (pm) Resonance wavelength shifts of TM and TE modes and the difference due to the change in outside refractive index in case of air surrounding medium n 0 = 1 1 10 8 − × 1 10 7 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × 1 10 0 × The change in outside refractive index Resonance wavelength shift (pm) TM TE Difference 63 4.3 Experimental Demonstration The experimental set up is shown in figure 4.4. Figure 4.4: Experiment setup. The 1550 nm laser light passes through a polarization controller and is butt coupled into a polymer waveguide that is vertically coupled to high Q fused silica resonator disc. With the waveguide design to satisfy the phase matching condition and the coupling mechanical setup, we can excite the fundamental WGM near the edge of disc resonator [29]. The output is also butt coupled to a single mode fiber and leads to a photo detector. The polarization controller is adjusted to observe simultaneously both the TE and TM polarization resonances. We measured the differential frequency shift between the set of TE and TM modes with the same angular mode numbers to detect the 64 temperature fluctuation or a change in the surrounding medium refractive index. In order to get a good measure of the temperature, all components were located inside a box to avoid temperature fluctuations of the room air (we did not actively stabilize the temperature). The measurement system consists of a ramp signal function generator to sweep the laser frequency, an oscilloscope and a computer to observe and record the transmission spectrum. Figure 4.5: Thorlabs Manual Fiber Polarization Controller FPC560. It is difficult to accurately measure the frequency of the coupling peaks in the transmission spectrum because the minima of these resonance peaks are rounded. Normally, in many published research article, the range for measurement resolution is about 1/15 to 1/50 of linewidth of output spectrum. With this range we can only resolve the change in refractive index of 10 -5 RIU as in theoretical plot. The low resolution of measurement system due to some reasons as: laser power is normally in mW range then the output spectrum in µW or nW due to the loss in coupling and resonator itself, the 65 Figure 4.6: The Transmission Output Spectrum by adjusting the positions of Thorlabs Manual Fiber Polarization Controller to achieve different polarization coupling state to the resonator. Upper: TM mode was strongly coupled to WGM resonator. Lower: TE mode was strongly coupled to WGM resonator. 66 Figure 4.7: The Transmission Output Spectrum by adjusting the positions of Thorlabs Manual Fiber Polarization Controller to achieve different polarization coupling state to the resonator. Both TE and TM modes were coupled to WGM resonator by control the position of Polarization Controller. responsibility of photodetector in the range of 1 A/W (for example Thorlabs DET410), the oscilloscope with minimum detect of 1mV/div and the quality factor of the resonator itself. For example, if the signal provide by photdetector is not strong enough (specially of the region near the resonance peak where the slope of the change in transmission amplitude over the change of wavelength is very small) one may observe the flat portion at the resonance peaks which make the limit in the resolution to measure the wavelength shift. To overcome this problem one may need more power on laser or more sensitivity photodetector which may limit in the market. 67 Figure 4.8: The zoom in of transmission output spectrum to show a flat portion at the resonant peak area. In order to detect a very small shift in resonance wavelength we need to propose a method of measurement with very high resolution. One useful approach is to use a lock-in amplifier to obtain the derivative of the transmission signal and to track its zero crossing point. As consequence of the relatively large slope and crossing zero at one point of the derivative signal, small temperature fluctuations or small variations in the outside refractive index can thereby be transformed into a detectable change. In principal, if we modulate the laser frequency by adding a small sinusoidal component to laser injection current, then laser frequency (f is time modulated around some average ) given by: (4.11) where is average laser frequency ω is modulation frequency is the frequency modulation amplitude Output Transmitted Spectrum Q = 2*10^6 -2.00E-03 -1.00E-03 0.00E+00 1.00E-03 2.00E-03 3.00E-03 4.00E-03 0.00E+00 5.00E+07 1.00E+08 1.50E+08 2.00E+08 2.50E+08 Frequency Tuning (Hz) Output Transmission Intensity 0 f ) t sin( A f f FM ω + = 0 0 f FM A 68 At the transmitted output of the system waveguide coupled to high Q WGM resonator, we get: (4.12) We can expand in Taylor series about , and yield: (4.13) Now we can use this signal to apply into the input port of the Lock-in Amplifier We use the same small sinusoidal which modulate the laser as the reference signal and we get: (4.14) Both reference and input signals get to the mixer inside Lock-in Amplifier and at the output of the mixer, we have (4.15) The DC component will be extracted by the Lowpass Filter and (4.16) Thus the Lock-in Amplifier detection can measure the derivative of the response with respect to the parameter which is modulated. )) t sin( A f ( T ) f ( T FM ω + = 0 0 f ) t sin( A ) f ( T ) f ( T ) f ( T FM ' ω 0 0 + = ) t sin( A ) t ( R FM φ ω + = .... ) t sin( ) t sin( A ) f ( T ) t sin( ) f ( T A ) f ( T ) t ( R FM ' FM + + + + = ⋅ ω φ ω φ ω 2 0 0 )] t cos( )[cos f ( T A ) t sin( ) f ( T A ' FM FM φ ω φ φ ω + + + + = 2 2 1 0 2 0 φ cos ) f ( T A V ' FM out 0 2 2 1 = 0 f 69 Figure 4.9: Upper: Output Transmission spectrum. Lower: A Lock-in Amplifier can obtain the derivative of the transmission output signal and help to measure the shift of resonant peaks by tracking its zero crossing points. 70 In the first experiment we demonstrated the reduced effect of temperature fluctuation on the difference between the TE and TM resonant frequencies. The Fused Silica resonator, radius R = 1.6 mm, was heated by illuminating the surface with a lamp and the shifts of each individual TE and TM modes and the difference frequency were recorded. The temperature was increased up to 5 o C above normal room temperature by increasing the power of the lamp and we allowed time in between each step for the system equilibration. The stability of the output temperature from a lamp at equilibration is ± 0.05 o C. For air surrounding the resonator, the resonant frequency shift of the TE mode, the TM mode, and the difference is shown in figure 4.10 as a function of temperature. The change in the difference frequency was less than the sensitivity of our measurement set up (±0.05pm) even at 5 o C change in temperature. However, it confirms the small effect of the temperature on the difference. We further demonstrated the measurement of a glucose solution by detecting the change in surrounding medium’s refractive index. The resonator’s surface was covered by a glucose solution with its concentration calculated by the weight percent (wt %) of glucose in de-ionized (DI) water. An aqueous solution of glucose has a refractive index that varies linearly with the concentration of glucose in water as 1.4 x 10 -3 RIU/wt % [30]. After each measurement with different glucose concentrations, we rinsed and cleaned the resonator disk with DI water. The output transmission spectrum was recorded after the system was stabilized. All measurements were done four times to prove repeatability. 71 Figure 4.10: The plot represents theoretical calculation vs. experimental data for the resonance shift of each individual TE and TM modes, and the change of the differential frequencies when the temperature is changed by 5 o C (the experimental system consists of Fused Silica resonator surrounded by air). Crosses, dots and triangles are corresponding to TE, TM resonance shifts and the differences respectively. The dash, dot and solid curves are the theory prediction of individual TM and TE resonant frequency shifts and the difference. The uncertainty of TM or TE resonance mode shift is ± 0.65pm. The observed changes in the differential frequency were within the accuracy of our measurement setup (± 0.05pm). Theoretically, the differential resonance shift is about 0.001 pm when temperature is change 5 o C. 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 0 1 2 3 4 5 6 Tem perature fluctuation ( o C) Resonance frequency shift (G Hz) TE and TM Differenti al Theoretical Prediction 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 0 1 2 3 4 5 6 Tem perature fluctuation ( o C) Resonance frequency shift (G Hz) TE and TM Differenti al Theoretical Prediction Oscilloscope Function Generator Coupling Waveguide Polarization controller Microdisk Oscilloscope Function Generator Coupling Waveguide Polarization controller Oscilloscope Detector Function Generator Coupling Waveguide Oscilloscope Detector Function Generator Coupling Waveguide Detector Function Generator Detector Detector Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Detector Function Generator Coupling Waveguide Polarization controller Microdisk k Oscilloscope Function Generator Coupling Waveguide Polarization controller Microdisk Oscilloscope Function Generator Coupling Waveguide Polarization controller Oscilloscope Detector Function Generator Coupling Waveguide Oscilloscope Detector Function Generator Coupling Waveguide Detector Function Generator Detector Detector Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Detector Function Generator Coupling Waveguide Polarization controller Microdisk k 72 We used the Glucose solution of 0.033 wt%, 0.5 wt% and 1 wt% correspond to the surrounding medium’s refractive indices of 1.32505, 1.3257, and 1.3264 respectively. We measured the shift of TE, TM resonance peaks and the difference through the transmission spectrum of the device and compare to the spectra of DI water (n = 1.325) as a reference point. Figure 4.11 shows the experimental average resonance frequency shifts and weighted error bars for TE, TM resonance peaks and the difference vs. the change in surrounding refractive index as calculated for the three Glucose solutions. Thermal fluctuation causes large uncertainty in resonant frequency shifts of individual TE and TM modes, while the uncertainty of correlated differential frequency measurement is significantly reduced, and by this means provides an enhanced detection limit [31]. Figure 4.12 shows the measurements of the glucose sensing and clearly illustrates the advantage of the using the difference frequency measurements. If one tracks the shift of either the TE or TM resonance, the data is inconsistent in that the shift for the 0.5 wt% solution is more than the shift for the 1.0 wt% solution. However, if one tracks the difference in the resonant frequencies the data is consistent. The inconsistency of the TE or TM shifts is likely due to a temperature drift during the measurement while the drift has little effect on the difference frequency measurement. 73 Figure 4.11: The plot illustrates the experimental data and the theory prediction. The changes in the outside refractive index were calculated for the three Glucose solutions. Oscilloscope Tunable Laser Source Detector Glucose Solution Syringe Function Generator Coupling Waveguide Polarization controller Microdisk Oscilloscope Tunable Laser Source Detector Glucose Solution Syringe Function Generator Coupling Waveguide Polarization controller Oscilloscope Tunable Laser Source Detector Glucose Solution Syringe Function Generator Coupling Waveguide Oscilloscope Tunable Laser Source Detector Glucose Solution Syringe Function Generator Coupling Waveguide Tunable Laser Source Detector Glucose Solution Syringe Function Generator Tunable Laser Source Detector Glucose Solution Syringe Tunable Laser Source Detector Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Tunable Laser Source Detector Glucose Solution Syringe Function Generator Coupling Waveguide Polarization controller Microdisk Microdisk 1 . 10 5 1 . 10 4 1 . 10 3 0.01 1 . 10 3 0.01 0.1 1 10 100 The change in outside refractive index Resonance wavelength shift (pm) Difference TM TE 1 . 10 5 1 . 10 4 1 . 10 3 0.01 1 . 10 3 0.01 0.1 1 10 100 The change in outside refractive index Resonance wavelength shift (pm) Difference TM TE 74 Figure 4.12: The output transmission spectrum when the detected samples are DI water, glucose concentrations of 0.5 wt% and 1 wt%. Based on the results, if a 1ºC temperature fluctuation is assumed, one would obtain a detection limit of 9.8 x 10 -5 RIU change in aqueous surrounding medium’s refractive index by using the differential measurement but only a detection limit of 5.6 x 10 -3 RIU change by using the conventional TE or TM measurement approach. As seen on Figure 4.1, the improvement in detection limit of a gas sensor would be even larger. The low detection limit can be achieved under ambient temperature fluctuation. Frequency deviation due to thermal expansion and thermoelastic fluctuation that are the common noises for both TE and TM modes are thusly eliminated [26]. 0 2 4 6 8 10 12 Frequency Detuning (GHz) Output Transmission Spectrum, a.u DI Water 0.5 wt% Glucose 1 wt% Glucose TE TM 0 2 4 6 8 10 12 Frequency Detuning (GHz) Output Transmission Spectrum, a.u DI Water 0.5 wt% Glucose 1 wt% Glucose 0 2 4 6 8 10 12 Frequency Detuning (GHz) Output Transmission Spectrum, a.u DI Water 0.5 wt% Glucose 1 wt% Glucose TE TM 75 4.4 Conclusion We demonstrate a novel method of measurement which employees the very small change of the differential frequency between TE and TM modes of WGM resonator due to temperature fluctuations. We have shown theoretically and experimentally that it reduces the influence of temperature on the measured results, as well as eliminating the common effect of thermal expansion and thermoelastic on the resonance peaks, thus significantly increasing the detection limit of sensors. In the demonstration, we calculated the detection limit improvement of about 2 orders of magnitude for an aqueous solution of glucose based on a 1ºC temperature fluctuation. With this new method and assuming a temperature controlled environment with ΔT ~ 0.01ºC, a gas sensor (n~1 on the outside) would have a detection limit of 10 -8 RIU. Compared to what is currently available the above represents a significant increase in detection limit and ability of detection under ambient temperature fluctuation. In addition, the technique can be applied to the active frequency stabilization schemes by using the differential frequency for both temperature measurement and temperature compensation. It is potentially attractive for frequency stabilization of laser or oscillator as well as other kinds of frequency references [26]. 76 4.5 Chapter References [1] C. Y. Chao, W. Fung, and L.J. Guo, “Polymer microring resonators for biochemical sensing applications,” IEEE Journal of Selected Topics Quantum Electronics, vol. 12, no. 1, pp. 134-142, 2006. [2] F. Vollmer, D. Braun, A. Libchaber, M. Khoshsima, I. Teraoka, and S. Arnold, “Protein detection by optical shift of a resonant microcavity,” Applied Physics Letters, vol. 80, pp. 4057-4059, 2002. [3] S. Arnold, M. Khoshsima, I. Teraoka, S. Holler, and F. Vollmer, “Shift of Whispering Gallery Modes in microsphere by protein adsorption,” Optics Letters, vol. 28, pp. 272-274, 2003. [4] N. M. Hanumegowda, C. J. Stica, B. C. Patel, I. White, and X. Fan, “Refractometric sensors based on microsphere resonators,” Applied Physics Letters, vol. 87, 2005. [5] A. Ksendzov and Y. Lin, “Integrated optics ring-resonator sensors for protein detection,” Optics Letters, vol. 30, pp. 3344-3346, 2005. [6] V. Zamora, A. Diez, M. V. Andres, and B. Gimeno, “Refractometric sensor based on whispering-gallery modes of thin capillaries,” Optics Express, vol. 15, pp. 12011-12016, 2007. [7] R. W. Boyd and J. E. Heebner, “Sensitive disk resonator photonic biosensor,” Applied Optics, vol. 40, pp. 5742-5747, 2001. [8] A. M. Armani, R. P. Kulkarni, S. E. Fraser, R. C. Flagan, and K. J. Vahala, “Label- Free, Single-Molecule Detection with Optical Microcavities,” Science, vol. 317, pp. 783-787, 2007. [9] A. A. Savchenko, A. B. Matsko, V. S. Ilchenko, N. Yu, and L. Maleki, “Whispering- Gallery-Mode resonators as frequency references. II. Stabilization,” Journal of the Optical Society of America B, vol. 24, pp. 2988-2997, 2007. [10] Melles Griot. Available: http://optics.mellesgriot.com/opguide/mp_3_2.htm [11] C. C. Lam, P.T. Leung, and K. Young, “Explicit asymptotic formulas for positions, widths and strengths of resonances in Mie scattering,” Journal of the Optical Society of America B, vol. 9, pp. 1585-1592, 1992. 77 [12] T. Le, A. A. Savchenkov, H. Tazawa, W. H. Steier, and L. Maleki, “Polymer optical waveguide vertically coupled to high Q Whispering–Gallery resonators,” IEEE Photonics Technology Letters, vol. 18, pp. 859-861, 2006. [13] Z. Weissman, E. Brand, I. Tsimberov, D. Brook, S. Ruschin, “Mach-Zehnder type, evanescent wave sensor, using periodically segmented waveguide,” in Laser and Electro- Optics Society Annual Meeting, 1998, vol. 2, pp. 85-86. [14] J. L. Remo, “Reduced-noise-displacement measurements with a correlated differential photodiode sensors,” Applied Optics, vol. 36, no. 22, Aug. 1997. 78 Chapter 5 Integrated Microfluidic/ Whispering Gallery Mode Based Sensor Systems for Chemical and Biological Analysis 5.1 Microfluidics and Its Advantages The microfluidics system has rapidly developed from an early single channel device to a complex analysis system [1, 2]. Several reviews of the microfluidics technologies have been published [3, 4]. The earliest microfluidic devices demonstrated that fluidic components could be miniaturized and integrated together, which lead to the idea that one could fit an entire “lab on a chip”: in much the same way that a microelectronic circuit is an entire computer on a chip. Since then, there has been a tremendous interest in harnessing the full potential of this approach, and, consequently, the development of a number of countless microfluidics devices and fabrication methods. The use of microfluidics devices clinically has a number of significant advantages: the amount of analytes used is quite small because of the size volume of fluids within the microfluidics channels and the fabrications techniques used to construct a device are based on the matured technique and materials. Similar to that for microelectronics, microfluidics technologies enable the fabrication of highly integrated devices for performing several different functions on the same substrate chip. 79 Most of the microfluidics systems were fabricated by using glass, polymer, or a combination of both as materials; while the channels were constructed by applying laser ablation, chemical etching, plasma-enhanced chemical vapor deposition, or soft lithography techniques. A polymer based system is suitable for detection because it requires optically transparent substrates, good adherence, and good flexibility [5]. On the other hand, the glass based system is useful for a system that needs mechanical stability and surface patternability, such as in nanofluidics, in which channels with rigid wall can be useful. In addition, using glass-polymer-glass sandwich configuration is an advantage because it relies on the use of a polymer layer as a bonding and working layer for lateral fluid confinements. Also, it is flexible and can be deformed around electrodes to provide a leak-free sealing with the patterned substrate and glass is used to compensate for the limit of durability as well as capability to withstand high pressures of polymer [6]. The low cost and the simplicity of PDMS fabrication lead to no specialized equipment or facilities. Once its molds are prepared, rapid prototyping enables devices can be made on short time scales with minimal expense. On other hand, devices fabricated from glass or silicon requires numerous processes such as chemical etching, RIE, thermal bonding, which takes time and effort. Moreover, the rapid growth of microfluidics technology is also due to the development in MEMS devices and biosensors as well as the need of portable clinical diagnostic devices for real time and on site detection, in order to eliminate time consuming laboratory analysis procedures. It will ultimately require the integration of sensing elements with electronic circuits and the handling sample system in which the 80 microfluidics channels is a strong candidate. With an integration of all the necessary elements to perform a specific biochemical analysis, these sensor systems, or micro total analysis systems, will show to have advantages in several areas including improved performance such as sensitivity and low detection limit, reduced consumption of sample, portability, and low cost fabrication. 5.2 Ultra High Q WGM Based Sensors and The Need of Practical Devices Ultra high Q WGM based sensors have many advantages, especially their high sensitivity. However, all of them are only demonstrating in the laboratory. To bring them out of a lab, into a practical form, leads to many challenges. One of these challenges is how to handle the samples, which was often used by dropping fluidic from the needle of the syringe in Laboratory’s experiments. Recently, there are demonstrations of an on-chip microfluidics using WGM resonators as sensing devices; nevertheless, the sensitivities are not quite as high because of the low Q of WGM resonators in the system [7, 8]. Our High Q crystalline WGM resonators have demonstrated as optical biosensors with ultra high sensitivity, real time, label-free, and compact. They also can significantly reduce the effect of temperature fluctuation by our novel method of using the differential TE and TM resonant modes measurement to dramatically improve the detection limit [9]. In addition, the materials of our WGM resonators are crystal Fused Silica or CaF 2 , of which, are strongly compatible with popular materials that used in microfluidics fabrication. Therefore, the small samples volume requirement, the device size and cost reduction, and 81 the superior advantage of combining with high Q WGM base sensing system are definitely major driving forces for the commitment to a high Q WGM based sensor with microfluidics technology. For utilizing the microfluidics technology into the high sensitivity practical sensors, we need a practical structure to integrate the high Q WGM resonators and form the channels for flowing fluidics. We can use glass, or PDMS, as a material to satisfy two functions: defining the microfluidics channels and serving as the resonator’s holder. 5.3 Fabrication a Prototype of The Integration High Q WGM Resonator with Microfluidics Channel (Opto-Fluidics Device) We constructed the microfluidics channels in a glass substrate. In our prototype demonstration, we could use either the lithography chemical wet etching method or the diamond cutting method. The lithography chemical wet etching technique may take a long period of time because it needs to etch a large hole to cover the mm range resonator and etching depth about couples hundreds of µm for resonator holder part. We used a diamond drill bit to cut a partial cylinder (given the opening portion for coupling purpose) in a glass substrate with the depth about 300 µm. Then, we used the diamond cutting discs to create the channels with dimensions of 200 µm wide and 100 µm depth. All processes take about under 30 minutes in the lab. In order to fabricate a resonator that can work in the above microfluidics channel to form a sensing system, we propose two different methods for the resonator’s fabrication. First, we can fabricate two glass cylinder spacers that have diameters smaller 82 than the resonator’s diameter that bond to top and bottom surfaces of the resonator by using UV epoxy. They serve two purposes: they hold the resonator to separate from the glass microfluidics structure and they form the fluidics channels around the edge of the resonator. Now we can align and bond the resonator to the microfluidics cell as in figure 5.1 and 5.2. The edge of the resonator came out of the microfluidics cell enough for prism, angle-polished fiber or waveguide coupling (about 5 µm). The other method is to shape the resonator in the process of polishing. Here, we used a diamond thin film to cut and shape the edge of the resonator into the cone shape, and polished a flat portion to form the resonator, as shown in the figure 5.3. Then, we can align and bond the resonator to the microfluidics cell in the same way as other method of fabrication. Finally, we cover the top structure with a glass slide and bond it to the system. The integration of high Q WGM resonator into microfluidics structure is illustrated as in figure 5.1 Figure 5.1: The construction of the Microfluidics WGM based sensing cell. WGM Resonator Spacer s Top Cover Glass Bottom Glass and Microfluidic Channels 83 Figure 5.2: High Q WGM resonator (fabricating with 2 spacers on top and bottom surfaces) integrated with microfluidics. 84 Figure 5.3: High Q WGM resonator (fabricating by polishing a cone shape at the edge of resonator) integrated with microfluidics. 85 5.4 Experimental Demonstration and Results In figure 5.4, we set up the experiment to demonstrate our concept. The 1550 nm laser light passes through a polarization controller. The input light, from the laser source through a single mode fiber, is butt coupled into a polymer waveguide. Both the fiber coupler and polymer waveguide were set up and aligned in the same platform which can be adjusted the positions by the help of a 3D stage. The microfluidics WGM based sensor was held in another 3D stage, and it can align to couple with the polymer waveguide. In this experiment setup, the polymer waveguide will laterally couple to WGM resonator. We have an option to use epoxy to glue the polymer waveguide permanently to the microfluidics channel to form a sensing cell. Then the output transmission light is gathered by a collimator lens and sent to the photodetector. The polarization controller is adjusted to simultaneously observe the TE and TM polarization resonances. We measured the differential frequency shift between the set of TE and TM modes with the same angular mode numbers to detect a change in the surrounding medium refractive index. Thermal fluctuation causes large uncertainty in resonant frequency shifts of individual TE and TM modes. While the uncertainty of correlated differential frequencies measurement is significantly reduced, and by this means provides an enhanced detection limit. 86 Figure 5.4: High Q WGM based microfluidics sensing: experiment setup. 3D Stages PhotoDetector Lens Polymer Waveguide Butt Coupling Fiber Opto-Fluidics Cell Lens Butt Coupling Fiber Polymer Waveguide Opto-Fluidics Cell 87 We demonstrated the measurement of a glucose solution by detecting the change in the surrounding medium’s refractive index. A small drop of Glucose solution was injected into the opto-fluidics device through the channels inlet. Surface tension and adhesion forces between liquid molecules and the glass substrate helped to keep the liquid covered to the edge of the resonator in the coupling region (the point of resonator disk contacted to the coupler). A glucose solution with its concentration calculated by the weight percent (wt. %) of glucose in Deionized (DI) water covered the resonator’s surface near the edge where the fundamental WGM locates. An aqueous solution of glucose has a refractive index that varies linearly with the concentration of glucose in water as 1.4 × 10 −3 RIU/wt. % [10]. After each measurement with different glucose concentrations, we rinsed the cell with DI water. The output transmission spectrum was recorded after the system was stabilized. We used the glucose solution of 1 wt. %, 1.5 wt.%, 2 wt. %, 2.5 wt. % and 3wt. %, corresponding to the surrounding medium’s refractive indices of 1.3264, 1.3271, 1.3278, 1.3285 and 1.3292, respectively. We measured the shift of the differential TE and TM resonance peaks through the transmission spectrum of the device and used the spectra of DI water (n = 1.325) as a reference point. Figure 5.5 shows the experimental resonance frequency shifts for differential TE and TM modes versus the change in the surrounding refractive index as calculated for the these glucose solutions. With this method and an assumed temperature controlled environment with ΔT ≈ 0.01 °C, it would have a detection limit of 10 −6 RIU. Compared to what is currently available, the above represents a significant increase in its detection limit and capability of detection under ambient temperature fluctuation. 88 Figure 5.5: Experimental data: the resonance wavelength shifts for detecting 1 wt.%, 1.5 wt.%, 2 wt.%, 2.5 wt.% and 3 wt.% Glucose solution and the theory prediction. The changes in the outside refractive index were calculated for glucose solutions and using the differential TE and TM mode measurement method. The advantage and attraction of high Q resonator, integrated with microfluidics sensing device, is the need of ultra small amount of sample for detection. We can fabricate and align the resonator so that the microfluidics channel is about a couple hundreds of micrometer around the edge of the resonator. Then the relationship between the requirement sample volumes versus the resonator’s radius, can be described as in figure 5.6. 89 Figure 5.6: The sample volume vs the radius of WGM resonators. With the microfluidics channel’s width from 100 µm to 500 µm and the resonator’s radius from 0.5 mm to 5 mm, the required volume of sample for detection is so small (in the range of 50 nL to 1.6µL). The low sample volume requirement (tens of microliters) is advantageous for clinical situations in which only a small amount of the biological fluid is available for assay. 90 5.5 Chapter References [1] D. J. Harrison, K. Fluri, K. Seiler, Z. Fan, C. S. Effenhauser, and A. Manz, “Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip,” Science, vol. 261, pp. 895-897, Aug. 1993. [2] M. J. Powers, K. Domansky, M. R. Kaazempur-Mofrad, A. Kalezi, A. Capitano, A. Upadhyaya, P. Kurzawski, K. E. Wack, D. B. Stolz, R. Kamm, L. G. Griffith, “A microfabricated Array Bioreactor for Perfused 3D Liver Culture,” Biotechnology and Bioengineering, vol. 78, no. 3, May 2002. [3] E. Verpoorte and N. F. de Rooij, “Microfluidics meets MEMS,” in Proceedings of the IEEE, Jun. 2003, vol. 91, no. 6, pp. 930-953. [4] T. D. Boone, Z. H. Fan, H. H. Hooper, A. J. Ricco, H. Tan, and S. J. Williams, “Plastic advances microfluidic devices,” Analytical Chemistry, vol. 74: pp. 78A-86A, Feb. 2002. [5] J. C. McDonald and G. M. Whitesides, “Poly(dimethylsiloxane) as a Material for Fabricating Microfluidic Devices,” Accounts of Chemical Research, vol. 35, no. 7, Jul. 2002. [6] A. Plecis and Y. Chen, “Fabrication of microfluidic devices based on glass-PDMS- glass technology,” Microelectronic Engineering, vol. 84, no. 5-8, pp. 1265-1269, 2007. [7] W. J. Roman, “On-Chip Microfluidic Integration of Ultra-High Quality Silicon Optical Microdisk Resonators for Lab-On-Chip Applications,” National Nanotechnology Infrastructure Network, REU Research Accomplishments, 2007. [8] A. Nitkowski and M. Lipson, “On-chip spectroscopy using compact silicon microring resonators integrated with microfluidic channels,” presented at the Conference on Lasers and Electro-Optics (CLEO), San Jose, CA, May 4, 2008. [9] T. Le, A. Savchenkov, N. Yu, L. Maleki, and W. H. Steier, “Optical resonant sensors: a method to reduce the effect of thermal drift,” Applied Optics, vol. 48, no. 3, Jan. 2009. [10] Z. Weissman, E. Brand, I. Tsimberov, D. Brook, and S. Ruschin, “Mach-Zehnder type, evanescent-wave sensor, using periodically segmented waveguide,” in Laser and Electro-Optics Society Annual Meeting, 1998, vol. 2, pp. 85-86.. 91 Chapter 6: Reaction Kinetics Simulation of Microfluidics Whispering Gallery Mode Resonator Based Sensors 6.1 Introduction WGM based sensors that operate without a need for labeling offer the exciting possibility of handheld device development, for example, quick quantification of biomarkers in a person’s blood. Rapid and cheap access to this personalized medical information is integral to the ultimate vision of more accurate and quantitative diagnosis and treatment of illnesses. However, microfluidics biosensing performances can be limited by the diffusion of the analytes near the sensing surface [1]. Molecules displacements are governed by diffusion that will affect practical time scale and sensing efficiency. To overcome this problem, the transport with convection will enhance the microfluidics sensing performance. Evidently, a key challenge in the development of WGM based incorporated with microfluidic is to understand its transport and reaction characteristics. The performance of microfluidic devices that rely on surface chemistry is controlled by the interplay between reaction kinetics and the rate of mass transfer to the reactive surfaces. Our work presents theoretical estimation and numerical simulation to explore the design 92 variables such as the role of flow rate, micro-channel geometry, the reaction depletion boundary layer and parameters of the antibody-antigen on its performance. A transport model, based on the convection, diffusion and chemical reaction processes, is developed that describes the optimized conditions for maximizing both the practical total reaction time and the efficiency of the system. 6.2 Theoretical Analysis of Microfluidic Assisted WGM Based Sensors Consider a WGM based sensor incorporated in a microfluidics device as in figure 6.1. The surface of the sensor (in our case will be the egde of WGM resonator where the evanescent field of the fundamental mode will be used to sense the analytes) is immobilized with receptors for the specific target molecules. The reaction occurs at the functionalized surface of the sensor. The reaction kinetics can be described by two-step process: 1) Mass-transport process: The analytes is transported by convection and diffusion from the bulk solution toward the reaction surface of sensor. [A] Bulk ⇒ [A] Surface 2) Reaction Process: The binding of specific analytes with immobilized receptors k a, k d [A] Surface + [B] ⇔ [AB] 93 Figure 6.1: Model of the sensing system is studied. Solution of target analyte with concentration c 0 flows with velocity v through our microfluidics channel of height h over our sensor based on WGM resonator. The edge surface of resonator is functionalized with b m receptors per unit area (mol/m 2 ). The polymer waveguide is used to couple the light in and out and it can be glued with microfluidics channel to form a sensing cell with one inlet and two outlets. where [A] Bulk is the analytes in the bulk with concentration c 0 [A] Surface is the analytes at the reaction surface with concentration c s [B] is the concentration of the receptors. [AB] is the concentration of analytes-receptors complex. k a is the association rate constant and k d is the dissociation rate constant Inlet Outlet 1 Outlet 2 Resonator with immobilized surface Polymer waveguide 94 Figure 6.2: The general model of microfluidics sensors which involved convection, diffusion and surface reaction processes in a planar geometry. h is the microfluidics channel’s height and L is the sensor’s surface length. The bulk analyte is affected by convection and diffusion process to diffuse to the reaction surface where the binding process will happen. 6.2.1 Mass-Transport (Convection-Diffusion) Process Diffusion process is of great importance in physics, chemistry and biology and it results in mixing of chemical substances when it moves materials from one point to another within a cell as a result of different concentration gradients. Purely diffusive transport is the random motion that involves random and uncorrelated steps of target molecules through the surrounding fluid. The time for a target molecule to diffusively reach a sensor surface thus scales statistically like the square of its distance h away. ! D = h 2 D (6.1) Usually, the sensor based on the purely diffusion will take a very long time to transport the target molecules to the reaction surface for detection. For example, take Reaction Convection Diffusion L h 95 D = 10 -11 m 2 /s, so the total time will be τ D = 25,000s (almost 7 hours) for the analytes to travel 500 µm distance to reach the sensor’s surface. Figure 6.3: The diffusion time as a function of analyte’s distance from the sensor reaction surface. Therefore, we need to add convective transport - the motion of target molecules along a fluid flow with velocity v to enhance the total detection time. Squires and his colleagues provide an excellent introduction to depletion layer effects in planar sensor performance and develop intuitive relationships between flow parameters and the transient response of those sensors [2]. The insights gained in that study provides guidance for understanding the parameters that govern the transient response of our non- planar WGM based sensors. Thus, depending on the geometry of the sensor cell which includes WGM resonator and microfluidics channel structures, we need to optimize the Mass-Transport process in order to achieve the practical time of detection and also improve the sensitivity of the sensor system. There are several parameters we should 0 1 10 4 − × 2 10 4 − × 3 10 4 − × 4 10 4 − × 5 10 4 − × 0 1 10 4 × 2 10 4 × 3 10 4 × Distance from sensor's surface (m) Diffusion Time (s) 96 consider in the process of optimization the system such as the Peclet number, the shear Peclet number, thickness of depletion layer, Damkohler number and the ideal reaction time. We can consider the pressure driven flow of the sample solution along channel with height h and width w. The velocity that the target is transported along the channel L s can be represented as: v= 6Q wh 3 z(h!z) (6.2) where Q is the volumetric flow rate, h is the height and w is the width of microfluidics channel. The Peclet number is the ratio of Diffusive Time over Convective Time, which often use to compare the affect of these processes. It shows that a target molecule can diffuse across the channel or to be swept downstream that same distance. P e = v*h D (6.3) v is the flow rate and D is diffusivity constant. When P e << 1, diffusion dominates and a depletion zone will propagate upstream a distance δ = h/P e . When P e >> 1, we also need to consider the shear Peclet number P es which is defined as: P es =6* L s h ! " # $ % & 2 *P e (6.4) where L s is the length of functionalized sensor’s capture surface along the direction of fluid flow. 97 The introduction of flow into the system will form the growth of the depletion layer indefinitely. A steady state situation is now attained with a steady value of depletion zone δ. For diffusive transport at the steady state, the total number of target arriving at the sensor’s surface per unit time is given by the diffusive flux J D = D*!n*h*w ! (6.5) where Δn is the concentration gradients. The diffusive flux will be precisely balanced by the target flux (molecule/time) delivered by convection J C =!n*Q (6.6) then the value of the steady-state depletion layer will be ! s = D*h*w Q = h P e (6.7) When P e << 1 diffusion dominate and the depleption zone will propagate upstream a distance. However, when P e >> 1 the depleption layer will also be depended on the geometry of the resonators and the microfluidics channels, and the thickness of depletion layer is ! ! L s P es 1 3 ! L s 6* L s h " # $ % & ' 2 * Q D*w " # $ $ % & ' ' 1 3 for P es >> 1 (6.8) ! ! L s P es 1 2 ! L s 6* L s h " # $ % & ' 2 * Q D*w " # $ $ % & ' ' 1 2 for P es << 1 (6.9) 98 Through these above factors, we need to optimize the flow velocity v based on the sensor’s geometry to improve the system. When P es >> 1, it will cause the decrease in the sensing efficiency, the target molecules will be swept downstream because it is not have enough time to diffuse to the sensor’s reaction surface. On the other hand, when P es << 1 diffusion wins because of the extremely slow flow rate, all target molecules are collected; however, the total time may not be practical. Figure 6.4: The depletion layer thickness as a function of flow velocity of bulk concentration through the microfluidics channel when P es >> 1. 6.2.2 Chemical Surface Reaction Assuming first-order Langmuir kinetics for simplicity, the surface concentration b(t) of receptors that are bound by target molecules obeys !b !t =k a c s (b m "b)"k d b (6.10) 99 where b m is the surface concentration of receptors on the sensor. Binding depends on the concentration of unbound sites (b m – b) and on target concentration c s at the sensor surface, whereas target molecules de-bind in proportion to the bound concentration. If convective and diffusive transport supplies target molecules much more quickly than reactions can bind them, then transport is ‘reaction limited’, we have c s !c o b(t) b m = c o (k d k a ) 1+c o (k d k a ) (1"e "(k a c o +k d )t ) (6.10) Let K D =k d k a (is Binding Affinity), the fraction of bound receptors in equilibrium, b eq , is b eq b m = c o K D 1+c 0 K D (6.11) If concentrated solutions ( ! c o K D >>1 or equivalently ! c o >>K D ) effectively saturate the sensor, whereas dilute solutions ( ! c o <<K D ) bind only a fraction and ! b eq "c o b m /K D <<b m (6.12) The sensor with area A will have ! N receptors =b m A receptors, and the number of receptor are bound in equilibrium is given by ! N receptors bound =b m Ac o /K D for the small sensors and dilute solutions. Less concentrated solution will result in a fraction of a target molecules bound at equilibrium. These above equations involve the time scale required for the sensor to equilibrate in the “reaction limited” case. This is the ideal total reaction time that presents an upper limit ! " R = (k a c o +k d ) #1 (6.13) 100 1 10 15 − × 1 10 14 − × 1 10 13 − × 1 10 12 − × 1 10 11 − × 1 10 10 − × 1 10 9 − × 1 10 14 − × 1 10 13 − × 1 10 12 − × 1 10 11 − × 1 10 10 − × 1 10 9 − × 1 10 8 − × 1 10 7 − × Concentration of Analyte (M) Equilibrium Concentration of Complex (mol/m^2) Figure 6.5: The saturated concentration of antibody-antigen complex as a function of the target concentration c s at the reaction surface. The red curve for Binding Affinity K D = 7.5*10 -14 M and the blue for K D = 10 -9 M. Figure 6.6: The ideal reaction time for different antibody plotted against analyte concentration. The simulation was calculated with different binding affinity parameters as in table 6.1. The red curve is for Streptavidin/Biotin, the blue curve is for Human C- reactive Protein, the black curve is for Immunoglobulin G and the pink curve is for Protein A33 Immunoassay. 1 10 14 − × 1 10 13 − × 1 10 12 − × 1 10 11 − × 1 10 10 − × 1 10 9 − × 1 10 8 − × 1 10 7 − × 0.1 1 10 100 1 10 3 × 1 10 4 × 1 10 5 × 1 10 6 × Concentration (M) Reaction Time (sec) 101 k a (M -1 *s -1) k d (s -1) K D (M) Streptavidin/Biotin [3] 4*10 7 3*10 -6 7.5*10 -14 Human C-reactive Protein [4] 10 7 2.6*10 -2 2.6*10 -9 Immunoglobulin G [4] 2.5*10 5 3*10 -4 1.2*10 -9 Protein A33 Immunoassay [5] 2.4*10 5 3.5*10 -3 1.46*10 -8 Table 6.1: The association rate constant, the disassociation rate constant and the binding affinity of various type of analytes. The Damköhler numbers (D a ) are dimensionless numbers used in chemical engineering to relate chemical reaction timescale to other phenomena occurring in a system. In our microfluidics model, we defined it as a ratio of reactive to transport flux. D a = k a b m ! D (6.14) If D a >> 1, mass transport is rate limiting and D a << 1, we have the reaction limited regime. If the analyte capture process is indeed “reaction limited”, we would expect that flow rate has no effect except as it determines the incubation time. In a “transport limited” case, the analyte capture process is enhanced with the convective flow. In our simulation model: k a = 4 x 10 4 (m 3 /(mol.s)), b m = 4.10 -8 (mol/m 2 ), δ = 100µm (which was calculated based on v = 20 µm/s), D = 7.4*10 -11 (m 2 /s), thus D a = 2.2*10 3 as shown in figure 6.7. In our model, the fluidics flow will significantly help to improve the total detection time. 102 Figure 6.7: Calculation of Damkohler number based on the thickness of depletion layer that is a function of the flow velocity. Red curve is for k a = 4 x 10 4 (m 3 /(mol.s)) and blue is for k a = 10 3 (m 3 /(mol.s)). Both curves used b m = 4*10 -8 mol/m 2 and D = 7.4*10 -11 m 2 /s. 6.3 Computational Simulations All computational simulations were completed with the commercially available finite-element method software COMSOL. Our study model consists of a WGM resonator as the sensor incorporated with microfluidics system. The edge surface of resonator is immobilized with biotin in order to capture the streptavidin that is target molecules. The initial conditions specified that the concentration of streptavidin was zero within the microfluidics channel and that no streptavidin was bound to the surface. First, the steady-state Navier–Stokes mode was solved to obtain the velocity profile within the microchannel. Then, using this solution, the transient convection– 1 10 7 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 0.1 1 10 100 1 10 3 × 1 10 4 × 1 10 5 × Depletion Layer Thickness (m) Damkohler Number 103 diffusion and surface binding modes were solved simultaneously (i.e. the weak-boundary mode of COMSOL solved the surface binding reaction and coupled it to the flux of antibody through the binding surface in the boundary conditions of the convection– diffusion mode). In the model, a 50 nM solution of streptavidin was introduced to the microchannel at a flow rate of 10µm/s. The diffusion coefficient for streptavidin was 7.4 *10 -11 m 2 /s [6]. The surface concentration of streptavidin binding sites is 4*10 -8 mol/m 2 which was based on experimentally measured values [7]. The k a = 4*10 7 M -1 s -1 and k d = 3*10 -6 s -1 were based on the experimentally measured equilibrium constants for the binding of streptavidin and biotin in solution [8]. The model assumed that the binding of streptavidin to the surface only occurred between one immobilized biotin molecule and one streptavidin binding pocket. Multiple biotin interactions for a single streptavidin molecule were not considered in the model. The viscosity 10 -3 kg m -1 s -1 and density 10 3 kg /m 3 of the solution were based on experimentally measured values for water at room temperature The microfluidics WGM based sensor system as shown in figure 6.1 which consists of 200 µm wide, 300 µm depth, 3mm length input channel that connected to a semi-circular chamber which contains WGM resonator. The chamber is a cylinder with 4 mm diameter and 300 µm tall with a part to be cut and glue the wafer that contains optical polymer waveguide as a mean of coupling the light in and out resonator. We also leave a gap between the semi-circular and the wafer for the fluidics outlet of the microfluidics system. The Fused Silica resonator is 3.2 mm diameter with taped edge shape which seat inside the chamber and contact with optical polymer waveguide. 104 We use COMSOL to characterize the flow field in the chamber with the inlet pressure defined relative to atmospheric pressure as a boundary setting condition. The flow field in the system is found by solving the incompressible Navier-Stokes equations. Typical the horizontal flow at the middle plan vertical cross section is shown as in Figure 6.8. The flow is quadratic at the input channel, and then spreads out into the chamber, giving a low flow field at the surface of WGM resonator. Figure 6.8: Vertical cross section of the flow field through microfluidics channel which incorporated with WGM resonator. 105 6.3.1 Effects of Flow Rate on The Reaction Kinetics Figure 6.9 (upper plot) shows the predicted transient process of the antigen– antibody binding at the reaction surface under different bulk concentration flow velocities. The curves are plotted as the average surface concentration of antigen– antibody complexes versus time for the microfluidics channel’s height of 400µm. For all the flow rates, the concentration of the antigen–antibody complexes increases over a period of time and thereafter reaches a plateau value that agrees very well with the theoretical equilibrium value (4*10 -8 mol/m 2 ). We defined the equilibrium binding time as the time at the beginning of the plateau. This equilibrium binding time increases with decreasing flow velocity because of insufficient mass transport of the analyte at low velocity. The worst kinetics happens at v = 0 since the replenishment of consumed analyte near the binding surface solely depends on the diffusion. However, further enhancement of reaction kinetics becomes less effect at high flow rates (i.e. v > 20 µm/s) and the reaction-limited zone is reached. The high flow velocity reduces the percentage of detected molecules because the target molecules are swept downstream before they can diffuse very far to reach the reactive surface. In microfluidic biochip applications, if the sample availability is limited and having a fast assay is less important, slow flow velocity should be used to deliver sample since the total consumption of the sample is proportional to the product of flow velocity and the equilibrium time. And it also provides the highest percentage of detected molecules. 106 Figure 6.9: Upper: The plot of surface concentration b(t) of receptors that are bound by target molecules as a function of total time which obtained by COMSOL simulation Lower: The total time which includes all processes (convection, diffusion and surface reaction) as a function of flow velocity. We simulated with different flow velocities: 2µm/s, 5µm/s, 10µm/s, 20µm/s, 50µm/s and 100µm/s 1 10 100 1000 0 20 40 60 80 100 120 Total Detection Time (Convection, Diffusion, Reaction) (s) Flow Rate (um/s) 107 6.3.2 Effects of Channel Height on The Reaction Kinetics Intuitionally, the higher microfluidics channel should have better reaction kinetics than the shallower one since the higher channel contains more analyte. We compared the binding reactions in different channel’s heights of 50µm, 100µm, 150µm, 200µm, 400µm and 1000 µm, at the same flow velocity v = 10 µm/s. Both Peclet number P e and shear Peclet number P es are functions of channel’s height h, so is the depletion layer thickness. Note that our model simulation has P es >> 1, thus according to equation (6.8) the depletion layer thickness is scale with P es -1/3 . When the sensor channel’s height is large, these depletion zones over the reaction surface area in these simulation cases did not change significantly. Moreover, they are relatively thinner compared to their channel’s heights which means mass transport is sufficient for the antigen–antibody binding. However, if the sensor channel’s heights are decreased, the depletion layer’s thickness becomes almost equal or even larger than the channel’s height. This is the case for channel’s height is smaller than 150 µm in our simulation model; consequently, the total detection time is slower. Therefore, for the fluid flow through our model of microfluidics WGM based sensing system, the shallow microfluidics channels are only preferred as long as there are no fabrication problem and clogging problem since they consume less analyte sample at the same flow velocity, however, they may cause slower detection time. 108 Figure 6.10: The total detection time with different channel heights: 50µm, 100µm, 150µm, 200µm, 400µm and 1mm which obtained by COMSOL simulation. 6.3.3 Effects of Bulk Concentration We also evaluated how strongly the concentration of antigen in the microfluidics channel affects the binding reaction. We compare the binding kinetics total detection time in different concentrations that vary from 10 nM to 500nM while the surface concentration of receptors is fixed at 4*10 -8 mol/m 2 . Figure 6.11 shows that it takes 25 times longer time for an analyte with a lower bulk concentration (10 nM) to reach equilibrium than an analyte with a higher concentration (500 nM). However, all the binding kinetics total detection times are in a reasonable practical range for on-field detection. The total time is in the minutes range for detecting nM concentration. However, it may become significant unpractical long detection time depending on the 0 10 20 30 40 50 60 70 80 90 100 0 200 400 600 800 1000 Total Detection Time (Convection, Diffusion, Reaction) (s) Channel Height (um) 109 binding affinity and how low the analyte concentration we want to detect. Comparing with case of purely diffusion, the total detection time is significantly enhanced by the introducing of convection process. Figure 6.11: The total time with different analyte concentrations: 10nM, 50nM, 100nM, 200nM, and 500nM which obtain by COMSOL simulation 6.4 Conclusion In the recent literature, many research groups developed vary methods to enhance the reaction kinetics. For example, analyte can be actively directed toward the sensor’s surface by the means of electrostatic field [9] or using the magnetic field gradients to control target molecules that are attached to magnetic particles [10]. Flowing sample is a 0 20 40 60 80 100 120 0 100 200 300 400 500 600 Total Detection Time (Convection, Diffusion, Reaction) (s) Concentration (nM) 110 simple method of increasing the reaction kinetics as in our theoretical analysis and computational simulation. We can also improve the enhancement by simply significantly decreasing the channel width (i.e couples micrometer); however, the pressure requires to maintain the volumetric flow rate may become unpractical and the alignment process can be difficult. Moreover, by the constrain of small volume analyte, we may think about reuse the sample by constructing the microfluidics channel such that the sample can be circular inside the channel or a structure can redirect the sample back to detection area. Some techniques could apply for our model such as Electrowetting or Electrothermal stirring [11, 12]. Using the transient COMSOL finite-element model developed in this report, we demonstrate that the microfluidics WGM based sensing system can be optimizing the design variables such as the role of flow velocity, micro-channel geometry, the reaction depletion boundary layer and parameters of the antibody-antigen to enhance the efficiency and practical detection time. The COMSOL finite-element model provides insights into the processes of optimization certain parameters and the results agreed well with our theoretical prediction. We develop an intuitive and practical understanding of analyte transport for our microfluidics WGM based sensing employing surface capture. 111 6.5 Chapter References [1] R. Shilton, M. K. Tan, L. Y. Yeo, and J. R. Friend, “Particle concentration and mixing in microdrops driven by focused surface acoustic waves,” Journal of Applied Physics, vol. 104, 2008. [2] T. M. Squires, R. J. Messinger, and S. R. Manalis, “Making it stick: convection, reaction and diffusion in surface-based biosensors,” Nature Biotechnology, vol. 26, no. 4, Apr. 2008. [3] L. S. Jung, et al., “Binding and dissociation kinetics of wild-type and mutant streptavidins on mixed biotin-containing alkylthiolate monolayers,” Langmuir, vol. 16, no. 24, pp. 9421-9432, 2000. [4] C. K. Yang, J. S. Chang, S. D. Chao, and K. C. Wu, “Effect of diffusion boundary layer on reaction kinetics of immunoassay in a biosensor,” Journal of Applied Physics, vol. 103, 2008. [5] B. Catimel, M. Nerrie, F. T. Lee, A. M. Scott, G. Ritter, S. Welt, L. J. Old, A. W. Burgess, E. C. Nice, “Kinetic analysis of the interaction between the monoclonal antibody A33 and its colonic epithelial antigen by the use of an optical biosensor - A comparison of immobilization strategies,” Journal of Chromatography A, vol. 776, no. 1, pp. 15-30, 1997. [6] J. Spinke, M. Liley, and F.-J. Schmitt, H.-J. Guder and L. Angermaier, W. KnoUa, “Molecular recognition at self-assembled monolayers: Optimization of surface functionalization,” Journal of Chemical. Physics, vol. 99, no. 9, Nov. 1993. [7] K. E. Nelson, L. Gamble, L. S. Jung, M. S. Boeckl, E. Naeemi, S. L. Golledge, T. Sasaki, D. G. Castner, C. T. Campbell, and P. S. Stayton, “Surface Characterization of Mixed Self-Assembled Monolayers Designed for Streptavidin Immobilization,” Langmuir, vol. 17, pp. 2807-2816, 2001. [8] P. C. Weber, J. J. Wendoloski, M. W. Pantoliano, and F. R. Salemmet, “Crystallographic and Thermodynamic Comparison of Natural and Synthetic Ligands Bound to Streptavidin,” Journal of the American Chemical Society, vol. 114, pp. 3197-3200, 1992. [9] M. J. Heller, A. H. Forster, E. Tu, “Active microelectronic chip devices which utilize controlled electrophoretic fields for multiplex DNA hybridization and other genomic applications,” Electrophoresis, vol. 21, no. 1, pp. 157-164, Jan. 2000. 112 [10] D. L. Graham, H. Ferreira, J. Bernardo, P. P. Freitas, and J. M. S. Cabral, “Single magnetic microsphere placement and detection on-chip using current line designs with integrated spin valve sensors: biotechnological applications,” Journal of Applied Physics, vol. 91, pp. 7786-7788, 2002. [11] F. Mugele and J. C. Baret, “Electrowetting: from basics to applications,” Journal of Physics: Condensed Matter, vol. 17, no. 28, Jul. 2005. [12] M. Sigurdson, D. Wang and C. D. Meinhart, “Electrothermal stirring for heterogeneous immunoassays,” Lab on a Chip, vol. 5, no. 12, pp. 1366-1373, 2005. 113 Chapter 7 Practical Integrated, Array and Portable Devices 7.1 Real-Time Ultra Sensitive Sensing System for Influenza A Virus Detection 7.1.1 Introduction The development of sensors for chemical and biological sensing is an extremely significant piece of research. It is very attractive because of its profitable nature towards helping improve human life such as bacteria and virus detection in medical diagnostics, screening of chemical compounds in drug discovery, food safety, environment monitoring, chemical and biological weapon detection for national security, and even organic compound detection for searching life on other planets. Two important characteristics of a sensor are selectivity which depends on the immobilized treatment of the detected surface, and sensitivity, that mainly depends on a type of transducer. The ultimate goal is to demonstrate a sensor with high selectivity and with the capability of an integrated optical compact structure combined with the potential for detection of several analytes simultaneously through the fabrication of multiples devices on a single chip and its ultra sensitivity to able to detect very low concentration or even a single particle or virus, i.e in the applications of biological weapon detection for national security, one may need the sensor which can detect one or several viruses for early prevention. One of the potential application for real-time, on field detection is an Influenza A virus portable sensor. Influenza is an acute respiratory disease that occurs annually, 114 causing fatality in the elderly and children, which consequently results in billions of dollars losses in business and productivity. The development of a biosensor technology will enable rapid and specific disease diagnosis on-site so that a clinician can quickly determine whether treatment is needed. Early identification of Influenza infection in poultry is critical for aiding in the control of outbreaks. Currently, three popular methods for Influenza A virus detection are: virus isolation, real-time reverse transcriptase polymerase chain reaction, and antigen capture immunoassays. Virus isolation is a sensitive technique, but typically requires 5–7 days for testing. Real-time reverse transcriptase polymerase chain reaction is a well-known detection method and is becoming more commonly available in veterinary diagnostic laboratories; however, complex processes such as requirement of expensive equipment and appropriate laboratory facilities [1, 2] make this approach inappropriate for real time detection and in situ analysis. Additionally, the antigen capture immunoassays, such as the commercially available Directigen [3] or Binax [4] tests, can provide rapid test results, but suffer from low sensitivity and relatively expensive [5]. With the advantages of high Q WGM resonators for a transducer system and a novel method of TE and TM mode measurement for eliminating temperature fluctuation, our technology proved experimentally that we can achieve very low detection limit. On other hand, the demonstration to integration of WGM based sensing system with micro fluidics technology will open a direction towards the development of a portable device with high sensitivity and ultra small sample requirement for real time detection. In which, it will lead to a rapid, on-field and high sensitive device to identify the Influenza A virus. 115 7.1.2 Theoretical Principal When analyzing a sample, we want to evaluate the concentration of specific bio- molecules. To achieve this, our sensor first needs to recognize the "target", take for example the Influenza A virus, from other proteins or molecules. The human body solved this problem a long time ago by using "antibodies". Antibodies are proteins that bind only to specific molecules, virus or bacteria. Having antibodies functionalized on a surface, the target molecules will only bind and stick near the surface as illustrated in figure 7.1. This is the basic process of "recognizing" a virus by a biosensor structure. Figure 7.1: Antibody-Antigen interaction: Illustrated the antibody (Y shape), the antibody's specific target (antigen; diamond shape), and an unspecific target (circle shape). Furthermore, specific analytes will bind to the antibody-coated resonator’s sensing surface, which is probed by the evanescent field of the guided mode, causing a corresponding resonant wavelength shift that is measured. Analysis of the resonant wavelength shift thus yields information on the amount of analytes. Resonator’s sensing surface 116 Consider the sensing surface of WGM resonator covering with a molecular layer of mass M. We have ! M ="*S*t (7.1) where ρ is the mass density of the molecular layer (g/cm 3 ) S is the sensing surface area t is the thickness of the molecular layer. We assume that the thickness of molecules layer is smaller than the evanescent field depth, which is normally about 400 nm in our prototype experiment, so that our resonant wavelength shifts are consistent with the theory. The WGM resonator will support the angular mode l to satisfy: ! l = 2"n eff R # (7.2) with n eff is the effective refractive index of resonator, R is its radius, and λ is the operating wavelength From equation (7.2), we can derive (by using ! "# "n eff = "# "M "M "t "t "n eff where ! "M "t =#S and ! "# "n eff = 2$R l ) ! "M = #S n eff $ %n eff %t & ' ( ) * + ,1 "$ (7.3) ! "# ="M 1 $S # n eff %n eff %t (7.4) 117 Consider a WGM resonator in our previous experiment with radius of 1.6 mm. If we immobilize the antibodies for recognizing the Influenza A virus on the circumference edge of resonator then the sensing surface area is about 0.5 mm 2 (2π x 1.6 mm x (50 x 10 -3 mm)). Influenza A virus has spherical shape with its diameter about 100nm and mass density of Influenza A layer ρ = 1.104 g/mol. We can calculate the change of effective refractive index over the change of molecular layer thickness ! "n eff "t is about 8.75*10 -4 /µm [6]. Δλ is the resonant wavelength shift 1 10 15 − × 1 10 14 − × 1 10 13 − × 1 10 12 − × 1 10 11 − × 1 10 10 − × 1 10 7 − × 1 10 6 − × 1 10 5 − × 1 10 4 − × 1 10 3 − × 1 10 2 − × 1 10 1 − × Mass of Analytes (g) Resonant wavelength shift (pm) Figure 7.2: The theoretical plot of Differential TE and TM modes resonant wavelength shift as a function of the mass of molecular layer on detecting surface of sensors. The calculation is based on Fused Silica WGM resonator with radius of 1.6 mm. 118 Virus or Protein Molecular Weight (g/mol) Mass of single virus or molecule (g) Influenza A 3 x 10 8 0.5 x 10 -15 Vaccinia 6 x 10 9 9.9 x 10 -15 Thyroglobulin 6.7 x 10 5 1.1 x 10 -18 Table 7.1: The Molecular Weight and the Mass of a single molecule or virus. With our ultra high Q factor of WGM resonator (10 7 for Fused Silica or 5 x 10 9 for CaF 2 ) plus the advantages of our differential TE and TM mode measurement, we can achieve the detectable resonant wavelength shift (assumed that the fluctuation of temperature in the range of 0.01 0 C) in order of 3 x 10 -4 pm [7]. Then according to figure 7.2 which is the plot of Differential TE and TM modes resonant wavelength shift as a function of the molecular layer’s mass, we can achieve the smallest detectable mass of molecular layer of 1.76*10 -12 g that equals to an amount of 3500 Influenza A viruses or detection limit of 3.5 pg/mm 2 . Even without parameters optimization, the sensitivity of our proposed Influenza A sensor is among the most sensitive sensors reported [8 - 10], and which is more sensitive than the commercially available Surface Plasmon Resonance sensors with detection limit of 10 pg/mm 2 [10]. The above estimation based on the geometry of WGM resonator, is used in our previous prototype experiment (Fused Silica, radius of 1.6 mm). The resonant wavelength shift signal follows a strong dependence on cavity curvature near the predicted 1/R scaling [11], providing a mechanism for increasing the sensitivity of sensing system by reducing the resonator radius and operating at a more favorable wavelength near 760 nm 119 where reduced water absorption. Thus, we can optimize the design and the dimension in resonator fabrication process to achieve higher sensitivity. Our options are to fabricate the smaller diameter resonator or design a process to functionalize a part of resonator’s circumference edge which will decrease the sensing surface area, lead to ultra low concentration of Influenza A virus or even can detect a single Influenza A virus. Moreover, most viruses have a characteristic size in the range of 20–200 nm [12], thus based on our theoretical estimation and data in table 7.1, we also constructed the sensing system to detect the Vaccinia Virus or a larger protein such as Thyroglobulin with ultra low detection limit. It has also opened a direction of construct an array sensor to detect couple target viruses at the same time, i.e Influenza A and Vaccinia detection. 7.1.3 Fabrication We can take advantage of our experimental demonstration of the high Q WGM resonator, integrated with microfluidics technology, to define the structure for our proposed sensor as illustrated in figure 7.3. In this case, we can use prism or optical polymer waveguide to couple the light in and out of the sensor. In addition, we can construct the integrated chip of an optical sensor, based on the polymer optical waveguide vertically coupled to the high Q resonator disk and detect the concentration of sample by the means of resonance wavelength shift from the transmission spectrum measured at the output of the polymer waveguide. 120 Figure 7.3: The proposed integrated chip of optical sensor based on polymer optical waveguide vertically coupled to high Q resonator disk. We can immobilize antibodies for Influenza A virus selectivity on the resonator sensing surface (the edge of WGM resonator) by using a method described in the reference [13]. Note that they performed all steps at room temperature. We will clean the resonator surface with a 2% Micro-90 solution (International Products Corporation, Burlington, NJ, USA) and rinsed it with deionized (DI) water, followed by immersing it in chromic acid for 30 min at 70–80 °C, rinsing with DI water and drying with a stream of clean air. Next, the resonator then were incubated with 2% MTS solution in anhydrous toluene for 1 hour under a dry nitrogen atmosphere, and rinsed with toluene and dried. The silanized resonator were then incubated with 2 mM MBS (prepared by dissolving 1 mg MBS in 0.3 ml anhydrous DMSO, then diluting to 1.5 ml with anhydrous ethanol) for 1 hour and rinsed with DI water. Next, 2 µl of antibody (1–5mg/ml) and 2 µl of goat IgG Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders Inlet Outlet WGM resonator Optical polymer waveguide Holder 121 (10 mg/ml) were applied to the sensing surface on resonator, and allowed to sit for 1 hour in a humid chamber. Finally, we rinsed the resonator with PBS and stored in SuperBlock® blocking buffer. With ultra high sensitivity, compact, real time and small sample requirement of our proposed sensor, it has potential for detecting a very low concentration of explosives, chemicals or biological agents. The capability of single particle or virus detection is feasible with appropriate design and fabrication. This is a big advantage for medical diagnosis or bioterrorism defense purposes where the sensor can quickly and accurately detect and identify in real time the presence of dangerous levels of the threat materials before it harms people and contaminates their surroundings. 7.2 Array WGM Based Sensors for Multiple Biomolecules Targets Detection An advance implementation of our proposed biomolecules sensor will be integrated into a miniature planar array of optical sensors for detecting specific biochemical in extremely small quantities and many biochemical species simultaneously. An array of this type would have an area of about 1 2 cm . Each element of the array would include an optical microresonator that would have a high value of the resonance quality factor. The surface of each microresonator would be derivatized to make it bind to the molecules of a species of interest. Such binding would introduce a measurable change in the optical properties of the microresonator, in our case, it is the shift of the resonance wavelength of these output transmission spectrum. Because each microresonator could be 122 derivatized for detection of a specific biochemical different from those of the other microresonators, it would be possible to detect multiple specific biochemicals by simultaneous or sequential interrogation of all the elements in the array. Integrated the high Q WGM resonators with optical polymer waveguide and microfludics technology will make the proposed sensing system one of the most promising practical sensing platform. The implemented structure can be either parallel or in series array as in figure 7.4. Ultimately we can construct an array of sensors in a combination form as illustrated in figure 7.5. Such interrogation would be effected by means of a grid of row and column polymer-based optical waveguides that would be made up of integral parts of a chip, in which the array would be fabricated. The row and column polymer-based optical waveguides would intersect at the elements of the array. At each intersection, the row and column waveguides would be optically coupled to one of the microresonators. The polymer-based waveguides would be connected via optical fibers to external light sources and photodetectors. One set of waveguides and fibers (e.g., the row waveguides and fibers) would couple light from the sources to the resonators; the other set of waveguides and fibers (e.g., the column waveguides and fibers) would couple light from the microresonators to the photodetectors. Each microresonator could be addressed individually by the row and column for the measurement of its optical transmission. Optionally, the chip could be fabricated so that each microresonator would lie inside a microwell, into which a microscopic liquid sample could be dispensed [14]. 123 Figure 7.4: Proposed series and parallel platforms array sensors for multiple targets detection Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders Analytes Inlet Analytes Outlet Light out to PhotoDetector Light in from Laser High Q resonators Optical Polymer Waveguides Fludic Channels Resonator Holders 124 Figure 7.5: Functionalized Optical Microresonators in an array would address optically via row and column optical polymer waveguides. It is a proposed combination platform of series and parallel detection array sensors. 7.3 Ultimate Sensitivity, Portable and On-Site Detection Whispering Gallery Mode Based Sensors It is suitable for the demand of our decade due to the need of miniaturized, portable, on-site detection devices with ultimate sensitivity. The high tolerance in temperature operating condition gives our proposed device the advantage of working on- site and opens the opportunity of constructing wireless network sensors with high sensitivity, covering a large area with minimum effect by a temperature variant. The wide range of potential applications includes fast medical diagnosis (detection bacteria, virus 125 or proteins), studying biomolecular interactions, environment uses (landmine detection, water quality, indoor air quality, food safety) and defense (bio war-fare agents). We propose to construct a handheld portable device for detection of analytes in low concentration and in a high tolerance environment temperature fluctuation. The device is based on our advantage technologies of: integrated high Q WGM resonator with optical polymer waveguide as sensing part, novel method of measurement by using differential frequency between TE and TM resonant modes for reducing effect of temperature fluctuation and providing the ability of ambient temperature tolerance operating. An electronics circuit will convert the optical spectrum to the user-friendly interfaces. 7.3.1 Theory Operation and Design Procedure of Electronics Readout Circuits We already experimentally demonstrate the microfluidics WGM based sensing cell to detect the analytes in previous chapters. However, we need an electronics circuit design that serves as the electronics read out mechanism for our portable sensing devices. It will take the output transmission signal (comes out of photodetector) and converting to an indicated number on the display that is proportional with the frequency different between TE and TM resonant peaks. Microfluidics WGM sensing system and electronics circuitry can be implement into a portable handheld sensor for real time on field detection. One can possible implement the wireless circuit into the proposed portable sensor and sends the results back to the center or even create a wireless sensing network. 126 The electronics implementation in our portable handheld device consists of three main parts as described in figure 7.6 Figure 7.6: Schematics of our proposed microfluidics WGM based portable sensor. 1) Electronics circuits to convert the differential wavelength between TE and TM resonance modes to trigger signal (START and STOP) 2) An oscillator circuit provides an ultra stable signal. In order to detect a very low analytes’ concentration, it corresponds to the shift of differential resonance modes. We need an oscillator with ultra high frequency accuracy, small temperature coefficient, and low power supply rejection ratio. The requirement is all above parameters much smaller than the thermal baseline. Typical, LC or RC 127 resonance circuit has a poor requirement parameter (10% error in frequency accuracy, 200ppm/ o C and 2500ppm/V); ceramic resonator achieve better, but quartz crystals are suitable for our high performance with parameters: about 0.005% error in frequency accuracy, temperature coefficient about 0.5 ppm/ o C, and power supply rejection ratio about 1ppm/V. 3) Counter and display circuit: receive the trigger signal from CONVERSION circuit to display information for users (i.e start counting and display when signal from low to high, then STOP counting and reset the counter when signal from high to low) 7.3.2 Circuit Implementation and Experimental Demonstration of a Practical Portable Sensor: We designed and built the circuit board as a schematic in figure 7.7 Figure 7.7: Schematics of electronics circuit. 128 Figure 7.7: Continued. 129 It can be briefly described as: the derivative of output transmission signal with the differential frequency between TE and TM peaks proportional to the concentration or refractive index change of surrounding medium is used as the input of electronics circuit. The circuit performs a couple functions: • Detection the two 0 crossing points of derivative of output transmission signal (corresponding to TE and TM mode resonant peaks) then convert it into a “START” and “STOP” control signal. • The circuit includes a stable oscillator to generate a 10KHz square wave signal (or maybe higher frequency for higher resolution of detection) • The “START’ and “STOP” control signal is used to control the counter and display circuit. When the control signal is in “START”, the counter counts the number of rising edge of the square wave from 10 KHz oscillator. When the control signal is in “STOP”, the counter stops counting. The indicated number on the display is the total number of rising edges in the time interval between “START” and “STOP”. It will be proportional with the different frequency between TE and TM mode resonant peaks in the output transmission signal. 130 Figure 7.8: Function blocks of electronics circuit. We used our microfluidics WGM based sensor, as these experiments described in chapter 5. The solution of Glucose will be used as the analyte we would like to detect. The setup is similar as our experiments in chapter 4. At the output of lock in amplifier, we have the derivative of a transmission signal with TE and TM resonant peaks that can be detected by the differential frequency measurement method in chapter 4. This signal, then, is provide for the electronics circuit. The differential frequency between TE and TM resonant peaks is converted into an indicated number which is display on the 4 digit LED as in the figure 7.10. START and STOP Signal 10 KHz Square Wave Generator Counter Display 131 Figure 7.9: The signals captured by oscilloscope shown: The “START” and “STOP” control points. The square waves are counted by counter between the time interval of two TE and TM mode resonant peaks, the derivative of output transmission comes from output of Lock in Amplifier Square wave signal from Oscillator circuit “START” “STOP” Derivative of output transmission signal comes from Lock in Amplifier 132 Figure 7.10: The electronics module and real time measurement display the total number of rising edges that were counted by counter. They are proportional to the differential frequency between TE and TM resonant peaks. The lower picture is the electronics circuit inside the module. As illustrating in figure 7.10, the LED displays 9102 which is corresponding to 9102 rising edges counted by the counter between the “START” and “STOP” commands from control signal. It is converted from 4.67 GHz frequency different between TE and TM modes resonant peaks. Then the sensitivity according to the change of 1 unit is 133 4.67 GHz / 9102 = 513KHz (or 4.11fm). It is sufficiency for our detection limit in the change of outside refractive index. We can always increase the sensitivity by using the higher frequency from oscillator circuit. (i.e 100 KHz or 1MHz). 134 7.4 Chapter References [1] A. Das and D. L. Suarez, “Development and bench validation of real-time reverse transcription polymerase chain reaction protocols for rapid detection of the subtypes H6, H9, and H11 of avian influenza viruses in experimental samples,” Journal of Veterinary Diagnostic Investigation, vol. 19, pp. 625-634, 2007. [2] E. Spackman, D. A. Senne, T. J. Myers, L. L. Bulaga, L. P. Garber, M. L. Perdue M, K. Lohman, L. T. Daum, D. L. Suarez, “Development of a Real-Time Reverse Transcriptase PCR Assay for Type A Influenza Virus and the Avian H5 and H7 Hemagglutinin Subtypes,” Journal of Clinical Microbiology, vol. 40, pp. 3256-3260, Sep. 2002. [3] Directigen Influenza Test Kits, Becton-Dickenson, Franklin Lakes, NJ. [4] BinaxNOW Influenza A & B, Inverness Medical Innovations, Inc. [5] P. R. Woolcock, C. J. Cardona, “Commercial Immunoassay Kits for the Detection of Influenza Virus Type A: Evaluation of Their Use with Poultry,” Avian Diseases, vol. 49, no. 4, pp. 477-481, Dec. 2005. [6] I. Teraoka and S. Arnold, “Theory of resonance shifts in TE and TM whispering gallery modes by nonradial perturbations for sensing applications,” Journal of the Optical Society of America B, vol. 23, no. 7, Jul. 2006. [7] T. Le, A. Savchenkov, N. Yu, L. Maleki, and W. H. Steier, “Optical resonant sensors: a method to reduce the effect of thermal drift,” Applied Optics, vol. 48, no. 3, Jan. 2009. [8] H. Li and X. Fan, “Characterization of sensing capability of optofluidic ring resonator biosensors,” Applied Physics Letters, vol. 97, 2010. [9] A. Brandenburg, “Differential refractometry by an integrated-optical Young interferometer,” Sensors and Actuators B: Chemical, vol. 39, pp. 266-271, 1997. [10] U. Jonsson, L. Fagerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, “Real- time biospecific interaction analysis using surface-plasmon resonance and a sensor chip technology,” Biotechniques, vol. 11, pp. 620-623, 1991. [11] S. Arnold, M. Khoshsima, I. Teraoka, S. Holler, and F. Vollmer, “Shift of whispering-gallery modes in microspheres by protein adsorption,” Optics Letters, vol. 28, no. 4, Feb. 2003. [12] C. A. Tidona, G. Darai, and C. Buuchen-Osmond, The Springer Index of Viruses. Berlin, Germany: Springer-Verlag, 2002. 135 [13] J. Xu, D. Suarez and D. S. Gottfried, “Detection of Avian Influenza virus using an interferometric biosensor,” Analytical and Bioanalytical Chemistry, vol. 389, pp. 1193-1199, 2007. [14] V. Ilchenko, L. Maleki, Y. Lin, and T. Le, “Integrated Miniature Arrays of Optical Biomolecule Detectors,” NASA Tech Brief, NPO-43164, Jul. 2009. 136 Chapter 8 Conclusion and Future Research Directions This research presented mathematical analysis and experimental demonstrated a novel method of differential TE and TM mode measurement that is significantly eliminating the effect of temperature fluctuation in WGM resonators leads to high sensitivity and low detection limit sensing applications. It also has demonstrated a planar optical polymer waveguide, coupled to the high Q WGM resonators and integrating the system into microfludics technology, to open the ability of implementing a practical array or portable on-field sensing device. In theory, we derived numerical analyses and compared them with the experiments’ results: • Derived the effective refractive index of the WGM resonators and the polymer waveguides as a function of their material refractive index and dimension, in order to optimize their coupling efficiency. • Derived the effect of temperature fluctuation and the surrounding medium refractive index of the WGM resonator in the change of the resonance peaks of sensing system in both case TE and TM modes. We also derive the effect of temperature fluctuation and the surrounding medium refractive index of the WGM resonator in differential TE and TM modes to prove the effectiveness of our novel method in eliminating the effect of temperature fluctuation, leading to ultra high sensitivity in detection of the sensing system. 137 • Modeled the microfluidics WGM resonator based sensors with immobilized sensing surface receptors for the specific target molecules detection and used this novel platform to study the binding kinetics of the biotin-streptavidin system. • Developed and investigated the flow-enhanced in our novel microfluidics sensor’s model which included the convection, diffusion and chemical reaction processes. Optimized platform’s parameters for maximizing both the practical total reaction time and the efficiency of the system. We experimentally demonstrated: • The fabrication of planar optical polymer waveguides with the effective refractive index matching with the effective refractive index of WGM resonator disk to perform efficient coupling. • The fabrication technique to achieve high Quality factor in WGM resonators. We demonstrated the technique in both materials Fused Silica and CaF 2 with the Quality factor of 2* 10 7 and 3*10 9 , respectively. • Optical polymer waveguide couples to high Q WGM resonator with efficiently critical coupling achievement. • The effect of temperature fluctuation and the effect of changing refractive index of glucose solution (as outside medium demonstration) on the shift of resonance peaks in both TE and TM modes. Experiment results shown that the temperature fluctuation effect on sensor sensitivity is significantly reduced by the use of our novel method of differential TE and TM mode measurement. 138 • The fabrication of microfludics channels in glass substrates by diamond cutting method and the integration of microfludics technology with high Q WGM resonators. We also demonstrated glucose sensing by using this integrated system. Combining this integrated sensing system with novel method of differential TE and TM mode measurement, we showed the ability of high sensitivity and low detection limits. The future directions of research and development in this study have so much potential. Following the investigations described in this proposal, a number of projects could be taken up. Some future research directions could provide the next steps along the path to the practical and widely applicable sensing systems. • The novel method of differential TE and TM mode measurement poses an extremely high tolerance with thermal effect ability that leads to low detection limit; however, it did not reach the floor limit through the resolution of the supported instruments, yet. Thus we can optimize the technique, expanding the research to push to a lower detection limit by exploring or combining with other options such as coating the micro resonators with negative refractive index coefficient materials, or with a fabrication technique to achieve the “identical” resonators (for reference method), or by fabricating the resonator with material which has the thermal expansion coefficient to cancel the thermo optic coefficient at certain temperature. • Integrated the high Q WGM resonators with optical polymer waveguide and microfludics technology will create the most promising practical sensing 139 platforms. We can demonstrate the array sensor for real-time simultaneous detection. Planar optical polymer waveguide offer the readiness for coupling with other photonics elements such as laser, photo-detector, and electronic circuits for an on-field, real time portable device, even with wireless transfer capacity or ultimately a lab-on-a-chip. 140 Bibliography A. M. Armani and K. J. Vahala, “Heavy water detection using ultra-high-Q microcavities,” Optics Letters, vol. 31, pp. 1896-1898, 2006. A. M. Armani, R. P. Kulkarni, S. E. Fraser, R. C. Flagan, and K. J. Vahala, “Label-Free, Single-Molecule Detection with Optical Microcavities,” Science, vol. 317, pp. 783-787, 2007. D. K. Armani, T. J. Kippenberg, S. M. Spillane, and K. J. Vahala, “Ultra-high-Q toroid microcavity on a chip,” Nature, vol. 421, pp. 925-928, 2003. S. Arnold, M. Khoshsima, I. Teraoka, S. Holler, and F. Vollmer, “Shift of whispering- gallery modes in microspheres by protein adsorption,” Optics Letters, vol. 28, no. 4, Feb. 2003. M. L. M. Balistreri, D. J. W. Klunder, F. C. Blom, A. Driessen, H. W. J. M. Hoekstra, J. P. Korterik, L. Kuipers, and N. F. Hulst, “Visualizing the whispering gallery modes in a cylindrical optical microcavity,” Optics Letters, vol. 24, pp.1829-1831, 1999. Directigen Influenza Test Kits, Becton-Dickenson, Franklin Lakes, NJ. B. Bhola, H. C. Song, H. Tazawa, and W. H. Steier, “Polymer microresonator strain sensors,” IEEE Photonics Technology Letters, vol. 17, no. 4, pp. 867-869, Apr. 2005. T. D. Boone, Z. H. Fan, H. H. Hooper, A. J. Ricco, H. Tan, and S. J. Williams, “Plastic advances microfluidic devices,” Analytical Chemistry, vol. 74: pp. 78A-86A, Feb. 2002. M. Borselli, T. J. Johnson, and O. Painter, “Beyond the Rayleigh scattering limit in high- Q silicon microdisk: Theory and experiment,” Optics Express, vol. 13, no. 5, pp. 1515- 1530, Mar. 2005. R. W. Boyd and J. E. Heebner, “Sensitive disk resonator photonic biosensor,” Applied Optics, vol. 40, pp. 5742-5747, 2001. V. B. Braginsky, M. L. Gorodetsky, and V. S. Ilchenko, “Quality-factor and nonlinear properties of optical whispering-gallery modes,” Physics Letters A, vol. 137, no. 7-8, pp. 393-397, May 1989. A. Brandenburg, “Differential refractometry by an integrated-optical Young interferometer,” Sensors and Actuators B: Chemical, vol. 39, pp. 266-271, 1997. 141 M. Cai, O. Painter, K. J. Vahala, and P. C. Sercel, “Fiber-coupled microsphere laser,” Optics Letters, vol. 25, pp.1430-1432, 2000. T. Carmon, L. Yang, and K. J. Vahala, “Dynamical thermal behavior and thermal self- stability of microcavities,” Optics Express, vol. 12, no. 20, pp. 4742-4750, 2004. B. Catimel, M. Nerrie, F. T. Lee, A. M. Scott, G. Ritter, S. Welt, L. J. Old, A. W. Burgess, E. C. Nice, “Kinetic analysis of the interaction between the monoclonal antibody A33 and its colonic epithelial antigen by the use of an optical biosensor - A comparison of immobilization strategies,” Journal of Chromatography A, vol. 776, no. 1, pp. 15-30, 1997. K. Chang, Handbook of Microwave and Optical Components. New York: Wiley, 2001. C. Y. Chao, W. Fung, and L.J. Guo, “Polymer microring resonators for biochemical sensing applications,” IEEE Journal of Selected Topics Quantum Electronics, vol. 12, no. 1, pp. 134-142, 2006. S. J. Choi, Z. Peng, Q. Yang, and P. D. Dapkus, “An eight channel demultiplexing switch array using vertically coupled active microdisk resonators,” IEEE Photonics Technology Letters, vol. 16, no. 11, pp. 2517-2519, Nov. 2004. L. Collot, V. Lefèvre-Seguin, M. Brune, J. M. Raimond, and S. Haroche, “Very High-Q Whispering-Gallery Mode Resonances Observed on Fused Silica Microspheres,” Europhysics Letters, vol. 23, no. 5, pp. 327-334, 1993. A. Das and D. L. Suarez, “Development and bench validation of real-time reverse transcription polymerase chain reaction protocols for rapid detection of the subtypes H6, H9, and H11 of avian influenza viruses in experimental samples,” Journal of Veterinary Diagnostic Investigation, vol. 19, pp. 625-634, 2007. V. V. Datsyuk, “Some characteristics of resonant electromagnetic modes in a dielectric sphere,” Applied Physics B: Lasers and Optics, vol. 54, no. 2, pp. 184-187, 1992. C. G. B Garrett, W. Kaiser, and W. L. Bond, “Stimulated emission into optical whispering gallery modes of spheres,” Physical Review, vol. 124, pp. 1807-1809, 1961. M. L. Gorodetsky, A. A. Savchenkov, and V. S. Ilchenko, “Ultimate Q of optical microsphere resonators,” Optics Letters, vol. 21, no. 7, pp. 453-455, 1996. D. L. Graham, H. Ferreira, J. Bernardo, P. P. Freitas, and J. M. S. Cabral, “Single magnetic microsphere placement and detection on-chip using current line designs with integrated spin valve sensors: biotechnological applications,” Journal of Applied Physics, vol. 91, pp. 7786-7788, 2002. 142 N. M. Hanumegowda, C. J. Stica, B. C. Patel, I. White and X. Fan, “Refractometric sensors based on microsphere resonators,” Applied Physics Letters, vol. 87, 2005. D. J. Harrison, K. Fluri, K. Seiler, Z. Fan, C. S. Effenhauser, and A. Manz, “Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip,” Science, vol. 261, pp. 895-897, Aug. 1993. M. J. Heller, A. H. Forster, E. Tu, “Active microelectronic chip devices which utilize controlled electrophoretic fields for multiplex DNA hybridization and other genomic applications,” Electrophoresis, vol. 21, no. 1, pp. 157-164, Jan. 2000. V. S. Ilchenko, A. A. Savchenkov, A. B. Matsko, and L. Maleki, “Nonlinear optics and crystalline whispering gallery mode cavities,” Physical Review Letters, vol. 92, 2004. V. Ilchenko, L. Maleki, Y. Lin, and T. Le, “Integrated Miniature Arrays of Optical Biomolecule Detectors,” NASA Tech Brief, NPO-43164, Jul. 2009. BinaxNOW ® Influenza A & B, Inverness Medical Innovations, Inc. U. Jonsson, L. Fagerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, “Real-time biospecific interaction analysis using surface-plasmon resonance and a sensor chip technology,” Biotechniques, vol. 11, pp. 620-623, 1991. L. S. Jung, et al., “Binding and dissociation kinetics of wild-type and mutant streptavidins on mixed biotin-containing alkylthiolate monolayers,” Langmuir, vol. 16, no. 24, pp. 9421-9432, 2000. A. Ksendzov and Y. Lin, “Integrated optics ring-resonator sensors for protein detection,” Optics Letters, vol. 30, pp. 3344-3346, 2005. J. C. Knight, H. S. T. Driver, R. J. Hutcheon, and G. N. Robertson, “Core resonance capillary fiber whispering gallery mode laser,” Optics Letters, vol. 17, pp. 1280-1282, 1992. C. C. Lam, P.T. Leung, and K. Young, “Explicit asymptotic formulas for positions, widths and strengths of resonances in Mie scattering,” Journal of the Optical Society of America B, vol. 9, pp. 1585-1592, 1992. T. Le, A. A. Savchenkov, H. Tazawa, W. H. Steier, and L. Maleki, “Polymer optical waveguide vertically coupled to high Q Whispering–Gallery resonators,” IEEE Photonics Technology Letters, vol. 18, pp. 859-861, 2006. 143 T. Le, A. Savchenkov, N. Yu, L. Maleki, and W. H. Steier, “Optical resonant sensors: a method to reduce the effect of thermal drift,” Applied Optics, vol. 48, no. 3, Jan. 2009. K. K. Lee, D. R. Lim, L. C. Kimerling, J. Shin, and F. Cerrina, “Fabrication of ultralow loss Si/SiO 2 waveguides by roughness reduction,” Optics Letters, vol. 26, no. 23, pp. 1888-1890, 2001. H. Li and X. Fan, “Characterization of sensing capability of optofluidic ring resonator biosensors,” Applied Physics Letters, vol. 97, 2010. B. E. Little, S. T. Chu, W. Pan, D. Ripin, T. Kaneko, Y. Kokubun, and E. Ippen, “Vertically coupled glass microring resonator channel dropping filters,” IEEE Photonics Technology Letters, vol. 11, no. 2, pp. 215-217, Feb. 1999. B. E. Little, J. S. Foresi, G. Steinmeyer, E. R. Thoen, S. T. Chu, H. A. Haus, E. P. Ippen, L. C. Kimerling, and W. Greene, “Ultra-compact Si–SiO 2 microring resonator optical channel dropping filters,” IEEE Photonics Technology Letters, vol. 10, pp. 549-551, 1998. L. Matthies, P. Bellutta, M. McHenry,“Detecting water hazards for autonomous off-road navigation,” Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA, USA 91109, 2003. D. Marcuse, Theory of Dielectric Optical Waveguides. New York: Academic, 1974. J. C. McDonald and G. M. Whitesides, “Poly(dimethylsiloxane) as a Material for Fabricating Microfluidic Devices,” Accounts of Chemical Research, vol. 35, no. 7, Jul. 2002. Melles Griot. Available: http://optics.mellesgriot.com/opguide/mp_3_2.htm F. Mugele and J. C. Baret, “Electrowetting: from basics to applications,” Journal of Physics: Condensed Matter, vol. 17, no. 28, Jul. 2005. K. E. Nelson, L. Gamble, L. S. Jung, M. S. Boeckl, E. Naeemi, S. L. Golledge, T. Sasaki, D. G. Castner, C. T. Campbell, and P. S. Stayton, “Surface Characterization of Mixed Self-Assembled Monolayers Designed for Streptavidin Immobilization,” Langmuir, vol. 17, pp. 2807-2816, 2001. A. Nitkowski and M. Lipson, “On-chip spectroscopy using compact silicon microring resonators integrated with microfluidic channels,” presented at the Conference on Lasers and Electro-Optics (CLEO), San Jose, CA, May 4, 2008. 144 M. Oxborrow, “Traceable 2D finite-element simulation of the whispering-gallery modes of axisymmetric electromagnetic resonators,” IEEE Transactions on Microwave theory and Techniques, vol. 55, no. 6, pp. 1209-1218, Jun. 2007. A. Plecis and Y. Chen, “Fabrication of microfluidic devices based on glass-PDMS-glass technology,” Microelectronic Engineering, vol. 84, no. 5-8, pp. 1265-1269, 2007. M. J. Powers, K. Domansky, M. R. Kaazempur-Mofrad, A. Kalezi, A. Capitano, A. Upadhyaya, P. Kurzawski, K. E. Wack, D. B. Stolz, R. Kamm, L. G. Griffith, “A microfabricated Array Bioreactor for Perfused 3D Liver Culture,” Biotechnology and Bioengineering, vol. 78, no. 3, May 2002. H. Quan and Z. Guo, “Simulation of single transparent molecule interaction with an optical microcavity,” Nanotechnology, vol. 18, 2007. P. Rabiei, W. H. Steier, C. Zhang, and L. R. Dalton, “Polymer micro-ring filters and modulators,” Journal of Lightwave Technology, Vol. 20, pp. 1968-1975, 2002. J. W. S. Rayleigh, The Theory of Sound, Volume Two. New York: Dover, 1945. L. Rayleigh, “Further applications of Bessel’s functions of high order to the Whispering Gallery and applied problems,” Philosophical Magazine, vol. 27, pp. 100-109, 1914. L. Rayleigh, “The problem of the Whispering Gallery,” Philosophical Magazine, vol. 20, pp. 1001-1004, 1910. J. L. Remo, “Reduced-noise-displacement measurements with a correlated differential photodiode sensors,” Applied Optics, vol. 36, no. 22, Aug. 1997. W. J. Roman, “On-Chip Microfluidic Integration of Ultra-High Quality Silicon Optical Microdisk Resonators for Lab-On-Chip Applications,” National Nanotechnology Infrastructure Network, REU Research Accomplishments, 2007. V. Sandoghdar, F. Treussart, J. Hare, V. Lefevre-Seguin, J. M. Raimond, and S. Haroche, “Very low threshold whispering-gallery-mode microsphere laser,” Physical Review A, vol. 54, pp. R1777-R1780, 1996. A. A. Savchenkov, V. S. Ilchenko, A. B. Matsko, and L. Maleki, “KiloHertz optical resonances in dielectric crystal cavities,” Physical Review A, vol. 70, 2004. A. A. Savchenkov, A. B. Matsko, V. S. IIchenko, and L. Maleki, “Optical resonators with ten million finesse,” Optics Express, vol. 15, pp. 6768-6773, 2007. 145 A. A. Savchenkov, A. B. Matsko, M. Mohageg, and L. Maleki, “Ringdown spectroscopy of stimulated Raman scattering in a whispering gallery mode resonator,” Optics Letters, vol. 32, pp. 497-499, 2007. A. A. Savchenko, A. B. Matsko, V. S. Ilchenko, N. Yu, and L. Maleki, “Whispering- Gallery-Mode resonators as frequency references. II. Stabilization,” Journal of the Optical Society of America B, vol. 24, pp. 2988-2997, 2007. R. Shilton, M. K. Tan, L. Y. Yeo, and J. R. Friend, “Particle concentration and mixing in microdrops driven by focused surface acoustic waves,” Journal of Applied Physics, vol. 104, 2008. M. Sigurdson, D. Wang and C. D. Meinhart, “Electrothermal stirring for heterogeneous immunoassays,” Lab on a Chip, pp. 1366-1373, 2005. E. Spackman, D. A. Senne, T. J. Myers, L. L. Bulaga, L. P. Garber, M. L. Perdue M, K. Lohman, L. T. Daum, D. L. Suarez, “Development of a Real-Time Reverse Transcriptase PCR Assay for Type A Influenza Virus and the Avian H5 and H7 Hemagglutinin Subtypes,” Journal of Clinical Microbiology, vol. 40, pp. 3256-3260, Sep. 2002. J. Spinke, M. Liley, and F.-J. Schmitt, H.-J. Guder and L. Angermaier, W. KnoUa, “Molecular recognition at self-assembled monolayers: Optimization of surface functionalization,” Journal of Chemical. Physics, vol. 99, no. 9, Nov. 1993. T. M. Squires, R. J. Messinger, and S. R. Manalis, “Making it stick: convection, reaction and diffusion in surface-based biosensors,” Nature Biotechnology, vol. 26, no. 4, Apr. 2008. M. Sumetsky, R. S. Windeler, Y. Dulashko, and X. Fan, “Optical liquid ring resonator sensor,” Optics Express, vol. 15, no. 22, Oct. 2007. T. Tamir, Guided-Wave Optoelectronics (Springer Series in Electronics and Photonics). Springer-Verlag, 1990. I. Teraoka, S. Arnold, and F. Vollmer, “Perturbation approach to resonance shifts of whispering-gallery modes in a dielectric microsphere as a probe of a surrounding medium,” Journal of the Optical Society of America B, vol. 20, pp. 1937-1946, 2003. I. Teraoka and S. Arnold, “Theory of resonance shifts in TE and TM whispering gallery modes by nonradial perturbations for sensing applications,” Journal of the Optical Society of America B, vol. 23, no. 7, Jul. 2006. C. A. Tidona, G. Darai, and C. Buuchen-Osmond, The Springer Index of Viruses. Berlin, Germany: Springer-Verlag, 2002. 146 D. V. Tishinin, P. D. Dapkus, A. E. Bond, I. Kim, C. K. Lin, and J. O’Brien, “Vertical resonant couplers with precision coupling efficiency control fabricated by wafer bonding,” IEEE Photonics Technology Letters, vol. 11, no. 8, pp. 1003-1005, Aug. 1999. D. W. Vernooy, A. Furusawa, N. P Georgiades, V. S Ilchenko, and H. J. Kimble, “Cavity QED with high-Q whispering gallery modes,” Physical Review A, vol. 57, pp. R2292- R2296, 1998. D. W. Vernooy, V. S. Ilchenko, H. Mabuchi, E. W. Streed, and H. J. Kimble, “High-Q measurements of fused-silica microspheres in the near infrared,” Optics Letters, vol. 23, no. 4, pp. 247-249, 1998. E. Verpoorte and N. F. de Rooij, “Microfluidics meets MEMS,” in Proceedings of the IEEE, Jun. 2003, vol. 91, no. 6, pp. 930-953. T. Vo-Dinh, L. Allain, “Biosensors for Medical Applications”, in Biomedical Photonics Handbook, T. Vo-Dinh, Ed. CRC Press, 2003. T. Vo-Dinh, B. M. Cullum, “Fluorescence Spectroscopy for Biomedical Diagnostics”, in Biomedical Photonics Handbook, T. Vo-Dinh, Ed. CRC Press, 2003. F. Vollmer, and S. Arnold, “Whispering-gallery-mode biosensing: label-free detection down to single molecules,” Nature Methods, vol. 5, pp. 591-596, 2008. F. Vollmer et al., “Protein detection by optical shift of a resonant microcavity,” Applied Physics Letters, vol. 80, no. 21, pp. 4057-4059, May 2002. P. C. Weber, J. J. Wendoloski, M. W. Pantoliano, and F. R. Salemmet, “Crystallographic and Thermodynamic Comparison of Natural and Synthetic Ligands Bound to Streptavidin,” Journal of the American Chemical Society, vol. 114, pp. 3197-3200, 1992. Z. Weissman, E. Brand, I. Tsimberov, D. Brook, S. Ruschin, “Mach-Zehnder type, evanescent wave sensor, using periodically segmented waveguide,” in Laser and Electro- Optics Society Annual Meeting, 1998, vol. 2, pp. 85-86. P. R. Woolcock, C. J. Cardona, “Commercial Immunoassay Kits for the Detection of Influenza Virus Type A: Evaluation of Their Use with Poultry,” Avian Diseases, vol. 49, no. 4, pp. 477-481, Dec. 2005. J. Xu, D. Suarez and D. S. Gottfried, “Detection of Avian Influenza virus using an interferometric biosensor,” Analytical and Bioanalytical Chemistry, vol. 389, pp. 1193-1199, 2007. 147 C. K. Yang, J. S. Chang, S. D. Chao, and K. C. Wu, “Effect of diffusion boundary layer on reaction kinetics of immunoassay in a biosensor,” Journal of Applied Physics, vol. 103, 2008. A. Yariv, “Critical coupling and its control in optical waveguide-ring resonator systems,” IEEE Photonics Technology Letters, vol. 14, no. 4, Apr. 2002. A. Yariv, Optical Electronics, 4 th ed. Holt McDougal, 1991. A. Yariv, “Universal relations for coupling optical power between microresonator and dielectric waveguides,” Electronics Letters, vol. 36, no. 4, 2000. V. Zamora, A. Diez, M. V. Andres, and B. Gimeno, “Refractometric sensor based on whispering-gallery modes of thin capillaries,” Optics Express, vol. 15, pp. 12011-12016, 2007.

Abstract (if available)

Abstract This work is a study of a practical Whispering Gallery Mode based sensing device that consists of optical polymer waveguides integrated with high Q Whispering Gallery Mode (WGM) resonators that is incorporated with microfluidics technology. We proposed, developed and experimental demonstrated a method that solves the issue of integrating high Q WGM resonators into practical sensors for real time and portable on field detection. ❧ We demonstrated a novel implementation of a polymer optical waveguide coupled to the whispering gallery modes of a fused silica resonator disk. This is the first step in bringing high Q WGM disks closer to the integrated optics technology. We obtained near-critical coupling and a loaded Q = 1.2 x 10⁷. ❧ The use of a WGM resonator with an ultra-high quality factor Q is promising in highly sensitive, label-free, lab-on-a-chip sensor applications. However, one issue that causes the reduction of sensitivity and detection limits of the resonant sensor class is the fluctuation of temperature in the surrounding medium. We investigated a novel method of using the differential frequency of TE and TM modes to reduce the thermal noise baseline. Then, we studied the temperature dependence of the WGM based sensors and experimentally demonstrated the reduction of temperature fluctuation

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Analysis and design of cleaved edge lithium niobate waveguide resonators

PDF

Optical simulation and development of novel whispering gallery mode microresonators

PDF

Fabrication and characterization of yoroidal resonators for optical process improvement

PDF

Applications of optical microring and microdisk resonators as physical, chemical, and biological sensors

PDF

Surface-enhanced on-chip whispering gallery mode resonators for improved nonlinear optics

PDF

Whispering gallery mode resonators for frequency metrology applications

PDF

Development of novel optical materials for whispering gallery mode resonator for nonlinear optics

PDF

High-speed and reconfigurable all-optical signal processing for phase and amplitude modulated signals

PDF

Application of optical forces in microphotonic systems

PDF

Development of optical devices for applications in photonic integrated circuit and sensing

PDF

Advanced modulation, detection, and monitoring techniques for optical communication systems

PDF

Active integrated photonic devices in single crystal LiNbO₃ micro-platelets and a hybrid Si-LiNbO₃ platform

PDF

Developing improved silica materials and devices for integrated optics applications

PDF

All-dielectric polymer electro-optic receivers and sensors

PDF

Titania and hybrid titania - silica sol-gel thin films and their applications in integrated optical devices

PDF

All-optical signal processing toward reconfigurable optical networks

PDF

Biological and chemical detection using optical resonant cavities

PDF

Sweep-free Brillouin optical time-domain analyzer

PDF

Microring resonator based filters and modulators: optical coupling control and applications to digital communications

PDF

Development of organic-inorganic optical microcavities for studying polymer thin films