No description available.

No description available.

<p>Over the past few years various successful miniaturization attempts of External Cavity Semiconductor Lasers (ECSL) were published. They built upon the rich literature of ECSL configurations that were extensively analyzed and improved upon since the 1960s. This was merged with the microfabrication techniques of 3D structures based on MEMS technology. The main drive for miniaturizing such tunable lasers in the recent past was the huge potential for such devices in all optical networks specifically as signal sources that enable Wavelength Division Multiplexing (WDM).<br />This thesis compares the different configurations chosen to build tunable lasers using MOEMS technology. Our criteria of comparison include wavelength tuning range, side mode suppression, tuning speed and device dimensions. Designs based on the simple ECSL with a movable external mirror suffered from the tradeoff between tuning range and Side Mode Suppression SMS. To overcome this limitation most designs adopted grating based tuning using the Littrow or Littman/Metcalf configurations. These configurations allow for much better tuning results but don’t lend themselves easily to miniaturization. The grating based devices were bulky and quite complicated to realize.<br />We propose the adoption of the Zhu/Cassidy double external cavity configuration. It retains the simplicity of the single external mirror configuration along with the tuning range and the SMS of including multiple tuning elements. In its original form this configuration suffered from mode hopping within the tuning range. Thorough simulation, design and experimental evidence is presented in this work to show that by extending the configuration to allow full control over both optical tuning elements this drawback can be eliminated.<br />Our proposed design would reduce the form factor to < 300μm x 200μm x 200μm. The voltage required to tune through all the modes is < 40V and the resonant frequency of the mirror is in the 10s of MHz order of magnitude. When coupled with a multimode laser of a sufficiently broad lasing profile this setup should enable a tuning range > 72nm with a SMS >20dB.</p> / Master of Applied Science (MASc)

Master of Science / Department of Electrical and Computer Engineering / Steven Warren / The continuous and autonomous real-time monitoring of cattle state of health can provide major benefits for the U.S. livestock industry and lead to a higher quality beef product. Complete real-time monitoring could not only lead to earlier detection of disease in individual animals and reduce the spread of disease to a larger herd, but it could ultimately reduce the cost and frequency of on-site veterinary consultations. This thesis details a wearable device that is mounted on cattle to collect data from a network of internal and external sensors. In addition to the basic data collection, this thesis will describe the infrastructure to communicate these data sets to a central database for permanent storage and future analysis. Physiological, ambient environment, and physical activity data are acquired by the various sensors to give a good indication of the state of health of an animal wearing the device. The communication of data from internal sensors to an external wearable receiver is of particular interest since tissue is not an ideal medium for radio-frequency data transmission. Past research has attempted to use such links with little success due to large signal attenuation at high frequencies and a package that becomes much too large to be usable at low frequencies. As a result, a wireless communications method employing magnetic inductance at relatively low frequencies over short distances is described here.

Master of Science / Department of Electrical Engineering / William B. Kuhn / Designers of mixed-signal systems must understand coupling mechanisms at the system, PC board, package and integrated circuit levels to control crosstalk, and thereby minimize degradation of system performance. This research examines coupling mechanisms in a RF-targeted high-resistivity partially-depleted Silicon-on-Insulator (SOI) IC process and applying similar coupling mitigation strategies from higher levels of design, proposes techniques to reduce coupling between sub-circuits on-chip. A series of test structures was fabricated with the goal of understanding and reducing the electric and magnetic field coupling at frequencies up to C-Band. Electric field coupling through the active-layer and substrate of the SOI wafer is compared for a variety of isolation methods including use of deep-trench surrounds, blocking channel-stopper implant, blocking metal-fill layers and using substrate contact guard-rings. Magnetic coupling is examined for on-chip inductors utilizing counter-winding techniques, using metal shields above noisy circuits, and through the relationship between separation and the coupling coefficient. Finally, coupling between bond pads employing the most effective electric field isolation strategies is examined. Lumped element circuit models are developed to show how different coupling mitigation strategies perform. Major conclusions relative to substrate coupling are 1) substrates with resistivity 1 kΩ·cm or greater act largely as a high-K insulators at sufficiently high frequency, 2) compared to capacitive coupling paths through the substrate, coupling through metal-fill has little effect and 3) the use of substrate contact guard-rings in multi-ground domain designs can result in significant coupling between domains if proper isolation strategies such as the use of deep-trench surrounds are not employed. The electric field coupling, in general, is strongly dependent on the impedance of the active-layer and frequency, with isolation exceeding 80 dB below 100 MHz and relatively high coupling values of 40 dB or more at upper S-band frequencies, depending on the geometries and mitigation strategy used. Magnetic coupling was found to be a strong function of circuit separation and the height of metal shields above the circuits. Finally, bond pads utilizing substrate contact guard-rings resulted in the highest degree of isolation and the lowest pad load capacitance of the methods tested.

<p>Electroporation is a biological cell's natural reaction to strong electric fields, where transient pores are created in the cell membrane. While electroporation holds promise of being a safe and effective tool for enhancing molecular delivery in numerous medical applications, it remains largely confined to preclinical research and clinical trials due to an incomplete understanding of the exact mechanisms involved. Muscle fibers are an important delivery target, but traditional theoretical studies of electroporation ignore the individual fiber geometry, making it impossible to study the unique transverse and longitudinal effects from the pulse stimulus. In these long, thin muscle fibers, the total reaction of the fiber to the electric field is due to fundamentally different effects from the constituent longitudinal and transverse components of the electric field generated by the pulse stimulus. While effects from the transverse component have been studied to some degree, the effects from the longitudinal component have not been considered. </p><p>This study develops a model of electroporation and delivery of small molecules in muscle tissue that includes effects from both the transverse and longitudinal components of the electric field. First, an asymptotic model of electric potential in an individual muscle fiber is derived that separates the full 3D boundary value problem into transverse and a longitudinal problems. The transverse and longitudinal problems each have their own respective source functions: the new "transverse activating function" and the well known longitudinal activating function (AF). This separation enhances analysis of the different effects from these two AFs and drastically reduces computational intensity. Electroporation is added to the asymptotic fiber model, and simplified two-compartment mass transport equations are derived from the full 3D conservation of mass equations to allow simulation of molecular uptake due to diffusion and the electric field. Special emphasis is placed on choosing model geometry, electrical, and pulsing parameters that are in accordance with experiments that study electroporation-mediated delivery of small molecules in the skeletal muscle of small mammals.</p><p>Simulations reveal that for fibers close to the electrodes the transverse AF dominates, but for fibers far from the electrodes the longitudinal AF enhances uptake by as much as 2000%. However, on the macroscopic tissue level, the increase in uptake from the longitudinal AF is no more than 10%, given that fibers far from the electrodes contribute so little to the total uptake in the tissue. The mechanism underlying the smaller effect from the longitudinal AF is found to be unique to the process of electroporation itself. Electroporation occurs on the short time scale of polarization via the transverse AF, drastically increases membrane conductance, and effectively precludes further creation of pores from charging of the membrane via the longitudinal AF. The exact value of enhancement in uptake from the longitudinal AF is shown to depend on pulsing, membrane, and tissue parameters. Finally, simulation results reproduce qualitative, and in some cases quantitative, behavior of uptake observed in experiments.</p><p>Overall, percent increase in total tissue uptake from the longitudinal AF is on the order of experimental variability, and this study corroborates previous theoretical models that neglect the effects from the longitudinal AF. However, previous models neglect the longitudinal AF without explanation, while the asymptotic fiber model is able to detail the mechanisms involved. Mechanisms revealed by the model offer insight into interpreting experimental results and increasing efficiency of delivery protocols. The model also rigorously derives a new transverse AF based on individual fiber geometry, which affects the spatial distribution of uptake in tissue differently than predicting uptake based on the magnitude of the electric field, as used in many published models. Results of this study are strictly valid for transport of small molecules through small non-growing pores. For gene therapy applications the model must be extended to transport of large DNA molecules through large pores, which may alter the importance of the longitudinal AF. In broader terms, the asymptotic model also provides a new, computationally efficient tool that may be used in studying the effect of transverse and longitudinal components of the field for other types of membrane dynamics in muscle and nerves.</p> / Dissertation

This dissertation develops surface integral equations using constraint-based Helmholtz decompositions for electromagnetic modeling. This new approach is applied to the electric field integral equation (EFIE), and it incorporates a Helmholtz decomposition (HD) of the current. For this reason, the new formulation is referred to as the EFIE-hd. The HD of the current is accomplished herein via appropriate surface integral constraints, and leads to a stable linear system. This strategy provides accurate solutions for the electric and magnetic fields at both high and low frequencies, it allows for the use of a locally corrected Nyström (LCN) discretization method for the resulting formulation, it is compatible with the local global solution framework, and it can be used with non-conformal meshes. To address large-scale and complex electromagnetic problems, an overlapped localizing local-global (OL-LOGOS) factorization is used to factorize the system matrix obtained from an LCN discretization of the augmented EFIE (AEFIE). The OL-LOGOS algorithm provides good asymptotic performance and error control when used with the AEFIE. This application is used to demonstrate the importance of using a well-conditioned formulation to obtain efficient performance from the factorization algorithm.

Quarter-wave switched oscillators (SWOs) are an important technology for the generation of high-power, moderate bandwidth (mesoband) wave forms. The use of SWOs in high power microwave sources has been discussed for the past 10 years [1-6], but a detailed discussion of the design of this type of oscillators for particular waveforms has been lacking. In this dissertation I develop a design methodology for a realization of SWOs, also known as MATRIX oscillators in the scientific community.A key element in the design of SWOs is the self-breakdown switch, which is created by a large electric field. In order for the switch to close as expected from the design, it is essential to manage the electrostatic field distribution inside the oscillator during the charging time. This enforces geometric constraints on the shape of the conductors inside MATRIX. At the same time, the electrodynamic operation of MATRIX is dependent on the geometry of the structure. In order to generate a geometry that satisfies both the electrostatic and electrodynamic constraints, a new approach is developed to generate this geometry using the 2-D static solution of the Laplace equation, subject to a particular set of boundary conditions. These boundary conditions are manipulated to generate equipotential lines with specific dimensions that satisfy the electrodynamic constraints. Meanwhile, these equipotential lines naturally support an electrostatic field distribution that meets the requirements for the switch operation.To study the electrodynamic aspects of MATRIX, three different (but interrelated) numerical models are built. Depending on the assumptions made in each model, different information about the electrodynamic properties of the designed SWO are obtained. In addition, the agreement and consistency between the different models, validate and give confidence in the calculated results.Another important aspect of the design process is understanding the relationship between the geometric parameters of MATRIX and the output waveforms. Using the numerical models, the relationship between the dimensions of MATRIX and its calculated resonant parameters are studied. Finally, I present a comprehensive design methodology that generates the geometry of a MATRIX system from the desired specification then calculates the radiated waveform.

Lying between the microwave and far infrared (IR) regions, the "terahertz gap" is a relatively unexplored frequency band in the electromagnetic spectrum that exhibits a unique combination of properties from its neighbors. Like in IR, many materials have characteristic absorption spectra in the terahertz (THz) band, facilitating the spectroscopic "fingerprinting" of compounds such as drugs and explosives. In addition, non-polar dielectric materials such as clothing, paper, and plastic are transparent to THz, just as they are to microwaves and millimeter waves. These factors, combined with sub-millimeter wavelengths and non-ionizing energy levels, makes sensing in the THz band uniquely suited for many NDE applications. In a typical nondestructive test, the objective is to detect a feature of interest within the object and provide an accurate estimate of some geometrical property of the feature. Notable examples include the thickness of a pharmaceutical tablet coating layer or the 3D location, size, and shape of a flaw or defect in an integrated circuit. While the material properties of the object under test are often tightly controlled and are generally known a priori, many objects of interest exhibit irregular surface topographies such as varying degrees of curvature over the extent of their surfaces. Common THz pulsed imaging (TPI) methods originally developed for objects with planar surfaces have been adapted for objects with curved surfaces through use of mechanical scanning procedures in which measurements are taken at normal incidence over the extent of the surface. While effective, these methods often require expensive robotic arm assemblies, the cost and complexity of which would likely be prohibitive should a large volume of tests be needed to be carried out on a production line. This work presents a robust and efficient physics-based image processing approach based on the mature field of parabolic equation methods, common to undersea acoustics, seismology, and other areas of science and engineering. The method allows the generation of accurate 3D THz tomographic images of objects with irregular, non-planar surfaces using a simple planar scan geometry, thereby facilitating the integration of 3D THz imaging into mainstream NDE use.

Metamaterials are artificially constructed "materials" which are formed from arrays of engineered elements. By designing individual elements as well as their interactions, the propagation of electromagnetic (EM) waves within the structure can be manipulated, so that new responses can be realised which may not be found in nature. The subject of this research concerns the study of miniaturised elements with strong EM responses so that the constructed metamaterial can better approximate an ordinary low-loss material. The research involves designing miniaturised magnetic resonators operating in the microwave frequency range. A novel resonator prototypes, so-called “helical resonators”, have been successfully designed and fabricated whose physical sizes can fall below 1% of the free space wavelength at resonance. This contributes to a size reduction of 90% compared with previously published work. In addition, an analytical model has been developed, so that the resonance parameters of a helical resonator have explicit expressions. In particular, a constant optimal metallic fill ratio is demonstrated to exist, which can achieve a minimum resonant frequency and a maximum miniaturisation for any given outmost dimension of the element. The accuracy of the model has been verified by both simulation and experiment. The frequency responses of fabricated helical elements were measured using a vector network analyser and a pair of small loop non- resonant dipole probes, and the parameters were extracted using the phase frequency fit method which proves to have the best accuracy and robustness. The properties of a regular square array of helical resonators are subsequently investigated, which can be regarded as a two-dimensional metamaterial. A relevant analytical model has been developed, which characterises the array as an equivalent sheet with surface current distributions, rather than an artificial medium with finite thickness. The relation between the macroscopic EM fields and the small scale properties of individual helical resonators are then established. In particular, the helical resonators are observed to be inherently chiral, thus the assembled interlocking array exhibits dichroism. The transmission coefficients for the circular EM waves with two different polarisation states have been derived, which have been verified by simulation and measurement results as well. In addition, it has been theoretically demonstrated that the resonator elements and their spacings can be engineered, so that the circular EM wave with one particular polarisation state can be totally attenuated around the element resonance, while the other state suffers negligible attenuation. A quadratic relation between the optimal array spacing and the elements’ quality factor has been demonstrated.