This thesis summarises research on the improved design, integration and expansion of past cortex-like computer vision models, following biologically-inspired methodologies. By adopting early theories and algorithms as a building block, particular interest has been shown for algorithmic parameterisation, feature extraction, invariance properties and classification. Overall, the major original contributions of this thesis have been: 1. The incorporation of a salient feature-based method for semantic feature extraction and refinement in object recognition. 2. The design and integration of colour features coupled with the existing morphological-based features for efficient and improved biologically-inspired object recognition. 3. The introduction of the illumination invariance property with colour constancy methods under a biologically-inspired framework. 4. The development and investigation of rotation invariance methods to improve robustness and compensate for the lack of such a mechanism in the original models. 5. Adaptive Gabor filter design that captures texture information, enhancing the morphological description of objects in a visual scene and improving the overall classification performance. 6. Instigation of pioneering research on Spiking Neural Network classification for biologically-inspired vision. Most of the above contributions have also been presented in two journal publications and five conference papers. The system has been fully developed and tested in computers using MATLAB under a variety of image datasets either created for the purposes of this work or obtained from the public domain.

Clear plastic fluidic devices with ports for incorporating electrodes to enable electrochemiluminescence (ECL) measurements were prepared using a low-cost, desktop three-dimensional (3D) printer based on stereolithography. Electrodes consisted of 0.5 mm pencil graphite rods and 0.5 mm silver wires inserted into commercially available 1/4 in.-28 threaded fittings. A bioimaging system equipped with a CCD camera was used to measure ECL generated at electrodes and small arrays using 0.2 M phosphate buffer solutions containing tris(2,2′-bipyridyl)dichlororuthenium(II) hexahydrate ([Ru(bpy)3]2+) with 100 mM tri-n-propylamine (TPA) as the coreactant. ECL signals produced at pencil graphite working electrodes were linear with respect to [Ru(bpy)3]2+ concentration for 9-900 μM [Ru(bpy)3]2+. The detection limit was found to be 7 μM using the CCD camera with exposure time set at 10 s. Electrode-to-electrode ECL signals varied by ±7.5%. Device performance was further evaluated using pencil graphite electrodes coated with multilayer poly(diallyldimethylammonium chloride) (PDDA)/DNA films. In these experiments, ECL resulted from the reaction of [Ru(bpy)3]3+ with guanines of DNA. ECL produced at these thin-film electrodes was linear with respect to [Ru(bpy)3]2+ concentration from 180 to 800 μM. These studies provide the first demonstration of ECL measurements obtained using a 3D-printed closed-channel fluidic device platform. The affordable, high-resolution 3D printer used in these studies enables easy, fast, and adaptable prototyping of fluidic devices capable of incorporating electrodes for measuring ECL.

3D printing has gained substantial interest as an adaptable and low-cost technology for rapid prototyping and production of research tools owing to its fast design-to-object workflow (Fig. 1), ease of operation, and ability to fabricate relatively complex and intricate structures directly from computer-aided design (CAD) representations. Due to the advantages 3D printing offers over other more time-consuming and labor-intensive fabrication methods like photolithography, 3D printing has been especially helpful in the development and production of flow cells and other fluidic devices. 3D printing allows for complex channel geometries, and the complete structure, including ports for connecting commercially available tubing, may be prepared from a single CAD file. As a result of these conveniences, 3D-printed fluidic devices have recently emerged as effective candidates for research in sensing applications. In these studies, we demonstrate electrochemical immunoassays for the biomarker protein S100B, which has been related to conditions like skin cancer and brain injuries, based on 3D-printed flow-cells with modularly integrated electrodes. The fluidic devices in these studies are prepared from photocurable resin and feature channel dimensions of ~400 µm. The device design includes ports for interfacing the channel with commercial fittings and tubing for fluid delivery as well as an access point for the antibody-modified electrode. Sensing is accomplished through a sandwich-type electrochemical immunoassay strategy, leading to sensitive detection of S100B.

No description available.

This work presents membrane development applicable in nanofluidic devices. These membranes can also be termed suspended thin films, supported on two or more edges. I first discuss motivation and background for developing these structures. Then I derive the formative principles for nanofluidic systems. Following the derivation of the Navier-Stokes and Washburn equations, I discuss applying these theories to planar nanofluidic capillaries and finish the derivation by discussing the forces that drive liquid flow in nanochannels. I next discuss the membrane development process, starting with my work in static height traps, and develop the concept of analyzing nanoparticles using suspended membranes. After reviewing the lessons learned from the double-nanopore project I discuss developing an oxide layer tuned to the needs of a membrane and present the design of an adjustable membrane structure. Afterward, I discuss modeling and simulating the structure, and present a procedure for fabricating robust membranes. I then explain applying the membrane structure to form a nanofluidic pump and document the process for recording and analyzing the pumping characteristics for nanodevices. As part of the pump section I propose a theory and model for predicting the behavior of the pumps. I next present applying active membranes as nanoparticle traps. I document a quick-turn optical profilometry method for charicterizing the devices, then present experimental data involving trapping. Early results show that the device functions as a nanoparticle concentrator and may work well as a size-based trap for nanoparticles. I conclude by summarizing the main contributions made during my course of study and by providing supplemental material to guide future research.

Acoustic wave sensors have proven useful in many fields as primarily mass sensitive devices capable of responding to small environmental perturbations. The focus of this dissertation is the development of a new type of surface acoustic wave device with application to material property measurement, and biological and chemical sensing. This device is a combination of three independent acoustic wave devices with these waves propagated across the same area, while retaining independence of actuation and sensor function. The development of a complete sensor system, and its use and operation are presented for several example cases of chemical and biomarker sensing, and sample manipulation. These include experimental and theoretical studies for organic vapor sensing, biological moiety sensing, acoustic streaming to remove loosely bound material, and optimization of designs for these applications.

Spectral element based numerical solvers are developed to simulate electrokinetically driven flows for micro-fluidic applications. Based on these numerical solvers, basic phenomena and devices for electrokinetic applications in micro and nano flows are systematically studied. As a first application, flow and species transport control in a grooved micro-channel using local electrokinetic forces are studied. Locally applied electric fields, zeta potential patterned grooved surfaces, and geometry are manipulated to control mixed electroosmotic/pressure driven flow in the grooved micro-channel. The controlled flow pattern enables entrapment and release of prescribed amounts of scalar species in the grooves. As another application, hydrodynamic/ electrokinetic focusing in a micro-channel is studied. External electric field, flow rate of pressure driven flow, and geometry in the micro-channel are manipulated to obtain the focusing point, which led to determination of the electrophoretic mobility and (relative) concentration of dilute mixtures. This technique can be used to identify and detect species in dilute mixtures.

With the emergence of Matrix Assisted Laser Desorption/Ionization-Time-of-flight as a tool for diagnosis of diseases via proteomics, there is an increasing need for greater sensitivity. Analysis of peptides by MALDI-TOF-MS is affected by sample formulation and spotting onto a MALDI target. This dissertation investigates a novel MALDI sample preparation technique, Induction Based Fluidics (IBF), for depositing precise volumes (pL to nL) of samples onto the target. We have seen that while using IBF, the induced electric field accompanying deposition enhances matrix crystallization yielding smaller crystals with more homogeneity, as compared to conventional manual micropipette (MP) depositions. An investigation of the signal-to-noise (S/N) for IBF deposition of tryptic digested Bovine Serum Albumin (BSA) showed a significant improvement in the signal-to-noise ratio for 0.5 and 0.25 pmol/µL BSA sample compared to equivalent MP depositions. The S/N enhancement for IBF and MP depositions of BSA were studied using à-cyano-4-hydroycinnamic acid (CHCA) and 2,5-dihydroxybenzoic acid (DHB) matrices, and CHCA showed better results than DHB . The exciting results obtained by IBF prompted us to probe sample morphology more fully and to relate morphology to the detections level and hopefully, to increase the utility of MALDI-TOF-MS for detection of a larger range of peptides. Morphology results were correlated to sensitivity limits using both dispensing techniques. Because of dissimilar rates of evaporation, different or uneven deposition thickness, or crystal lattice morphology, discontinuous crystallization patterns were observed for MP depositions. However, IBF deposited samples occupied less planar area with uniform distribution of crystals, thereby reducing sample crystal heterogeneity and laborious hunt for a "sweet" or "hot" spot to produce high quality spectra. The application of IBF was extended to the tryptic digested BSA protein using peptide mass fingerprinting. IBF deposition resulted in a larger number of detectable peptides as well as higher sequence coverage as compared to equivalent MP depositions. In last few decades, advanced research and potential applications in the field of microelectronics have spurred interest in the development of reticulated doped polymer films. Bis (ethylenedioxy) tetrathiafulvalene (BEDO-TTF)/Polycarbonate (PC) films were synthesized and characterized for use in hand-held real time explosives sensors, capable of detecting nitro-based compounds (nitroaromatics, nitoamines and nitroesters), which are the main components of Improvised Explosive Devices or IEDs. Reticulated doped polymer films were prepared by exposing solid solutions of BEDO-TTF in PC to iodine to form conductive charge transfer complexes. The resulting films exhibited room temperature conductivities ranging from 6.33-90.4*10-5 S cm-1. The colored iodine complexes in the film were reduced by cyclic voltammetry yielding conductive, colorless, transparent films. Dielectric analysis (DEA) was used to probe relaxations in neat PC and BEDO-TTF/PC showed that BEDO-TTF plasticized the PC and decreased the glass transition temperature. Two secondary relaxations appeared in PC films, whereas the transitions merged in the doped film. DEA also revealed conductivity relaxations above 180°C, which were characterized by the electric modulus formalism and showed that BEDO-TTF increased the alternating current, (AC) conductivity in PC.

The field of microfluidics has the potential to provide a number of products to better everyday life, but is still not well understood. In previous research performed in the field, microfluidics has been shown to exhibit behavior different from what would be expected through normal pipe flow theory. While some research has shown that fluid flow through microchannels does conform to the theoretical flow mechanics, and thus can be predicted and understood through use of well-known relations; other research performed has indicated that fluid flow through microchannels experiences higher or lower pressure losses than would be expected with macro scale theory. This work strives to further explore and explain this anomaly by focusing on simple straight rectangular channels of varying hydraulic diameters from 24 µm to 88 µm, in order to form a more basic understanding for fluid flow in microchannels. Water was pumped through each of these channels at a number of different flow rates, and the static pressure was measured in two locations, a set length apart. The measured pressure loss over this length for each flow rate was then recorded and analyzed to provide relations between pressure loss and hydraulic diameter. Through the data obtained in this study, microfluidic flow of Reynolds numbers greater than 40 and in channels as small as 48 µm in diameter experienced pressure losses predicted from macroscale theory. Below these values, the data was more random, but still showed some conformance to theory. A clear relationship between measured pressure loss and hydraulic diameters over the entire range of channels was also found for two different flow rates. It is hoped that the data obtained will provide a better understanding of microfluidics and pave the way for potential applications to be realized.

Biomarkers are measurable indicators of health status or disease state that can be used for diagnosis and may help guide patient treatment strategies. Enzyme-linked immunosorbent assays (ELISA) and other many clinical techniques currently used for measuring biomarker proteins lack sensitivity, demand high analysis cost, are often not well-suited for measuring multiple biomarkers in a single sample, and require long analysis times. Here, we demonstrate simple, low-cost 3D-printed flow-through devices with integrated electrodes modified with gold nanoparticles (AuNPs) for electrochemical immunoassays of S100B, a biomarker protein related to conditions like skin cancer and brain injuries. Flow-through devices are fabricated from photocurable-resin using a desktop digital light processing (DLP) projector-based 3D printer to produce 500-800 µm square cross-sectional fluidic channels. Threaded ports at the ends and center of the channel are included in the device design for connecting commercially available fittings for fluid delivery and integrating low-cost graphite electrodes for electrochemical biosensing.