Due to the inability of intra-articular injuries to adequately self-heal, current therapies are largely focused on palliative care and restoration of joint function rather than true regeneration. Subsequently tissue engineering of chondral and osteochondral tissue constructs has emerged as a promising strategy for the repair of partial and full-thickness intra-articular defects. Unfortunately, the fabrication of large tissue constructs is plagued by poor nutrient transport to the interior of the tissue resulting in poor tissue growth and necrosis. Further, for the specific case of osteochondral grafts, the presence of two distinct tissue types offers additional challenges related to cell sourcing, scaffolding strategies, and bioprocessing. To overcome these constraints, this dissertation was focused on the development and validation of a microfluidic hydrogel platform which reduces nutrient transport limitations within an engineered tissue construct through a serpentine microfluidic network embedded within the developing tissue. To this end, a microfluidic hydrogel was designed to meet the nutrition requirements of a developing tissue and validated through the cultivation of chondral tissue constructs of clinically relevant thicknesses. Additionally, optimal bioprocessing conditions with respect to morphogen delivery and hydrodynamic loading were pursued for the production of bony and cartilaginous tissue from bone marrow derived mesenchymal stem cells. Finally, the optimal bioprocessing conditions were implemented within MSC laden microfluidic hydrogels to spatially engineer the matrix composition of a biphasic osteochondral graft through directed differentiation.

Cette thèse concerne la détection des métaux lourds dans l’environnement et en particulier dans les eaux de surface et les sous-produits de désinfection de l’eau potable. Les deux catégories de contaminations ont des propriétés différentes de sorte que deux méthodes correspondantes ont été dévéloppées : l’une est basée sur des capteurs moléculaires fluorescents mis en oeuvre dans un micro-dispositif, l’autre est basée sur une détection électrochimique. Deux capteurs moléculaires fluorescents, Rhod-5N et DPPS-PEG, et plusieurs dispositifs microfluidiques ont été fabriqués et appliqués pour la détection des ions de métaux lourds, Cd (II) et Hg (II),dans les eaux de surface. Une nouveau circuit en PMMA est fabriqué par ablation laser femtoseconde et testé pour la détection de Cd2+ avec le Rhod-5N. De plus, des améliorations de la performance des circuits microfluidiques ont été faites. Une nouvelle méthode de détermination sensible de cinq acides haloacétiques (AHAs) dans les d'eaux a été développée. Elle est basée sur l'extraction électromembranaire (EME) avant électrophorèse capillaire avec détection de conductivité sans contact à couplage capacitif (CE-C4D). / This thesis is aimed at environmental contaminations detection, mainly heavy metal ions in surface water and disinfection by-products (DBPs) in drinking water. The two categories of contaminations have different properties so that two correspondent methods were developed: one is based on fluorescent molecular sensors in a microfabricated device, the other one is based on conductive detection. Two fluorescent molecular sensors, Rhod-5N and DPPS-PEG, and several microfluidic devices were developed and applied for heavy metal ions Cd (II) and Hg (II) detection in surface water. A new microchip made of PMMA was fabricated by femtosecond laser ablation and tested for Cd (II) sensing based on a fluorescent molecular sensor Rhod-5N. Further more, some improvements of the performance of microfluidic chips were made. A novel method for sensitive determination of five priority haloacetic acids (HAAs) in water systems has been developed based on electromembrane extraction (EME) prior to capillary electrophoresis with capacitively coupled contactless conductivity detection (CE-C4D).

Micro-fluid mixing is an important aspect of many of the various micro-fluidic systems used in biochemical production, biomedical industries, micro-energy systems and some electronic devices. Typically, because of size constraints and laminar flow conditions, different fluids may only have the opportunity to mix by diffusion, which is extremely rate limited. Therefore, active or highly effective passive mixing techniques are often required. In this study, two pulsed injectors are used to actively enhance mixing in a high aspect ratio microchannel (125 ��m deep and 1 mm wide). The main channel has two adjacent flowing streams with 100% dye and 0% dye concentrations, respectively. Two injectors (125 ��m deep and 250 ��m wide) are located on separate sides of the channel, with one downstream 2 mm (equivalent to two main channel widths or eight injector widths) from the other. This results in an asymmetric mixing as the flow proceeds downstream. A dye solution is used to map local mixing throughout the channel by measuring concentration variations as a function of both space and time. The primary flow rates are varied from 0.01 to 0.20 ml/min (Reynolds numbers of 0.3 to 26.6), the injector flow rate ratios are varied from 0.125 to 2, and the pulsing frequencies are varied from 5 to 15 Hz. Images of the concentration variations within the channel are used to quantify mixing by calibrating the intensity of the image with the concentration of the dye solution. The degree of mixing (DoM) is used as a measure of quality and is defined based on the integration across the channel of the difference between the local concentration and the 50% concentration values. DoM is normalized by the 50% concentration value and subtracted from one to yield a parameter that varies from 0 (no mixing) to 1 (perfect mixing). It is shown that there is a high degree of repeatability of concentration distribution as a function of phase of the pulsing cycle. A mixing map is constructed over the range of variables tested which indicates an optimum set of flow and pulsing conditions needed to achieve maximum mixing in the main channel flow. The flow rate ratio between the injectors and main channel is found to be the most influential parameter on overall mixing. The highest DoM in the main channel was found to be 0.89. It is also noticed that improved mixing can occur at very low flow ratios under a unique set of primary flow and low frequency pulsing conditions. In general, there is an inverse relationship between primary flow rate and pulsing frequency to achieve better overall mixing. / Graduation date: 2003

Electro-transmittance of molecular junctions was characterized electrically and studied optically at 410nm and 532nm. Between 1kHz and 100kHz there was no qualitative difference between the control samples and the molecular junction samples, however there were difficulties with reproducibility of the quantitative behaviour, so no hard conclusions could be drawn. A microfluidic capacitor device was designed and fabricated to study the electrical double layer, using standard microfabrication techniques. A complimentary flux corrected transport simulation was written using the same experimental geometry and the results of this study found qualitative agreement between the simulation and experiment. The experiment produced results about the concentration dependence of the double layer formation time which allows an estimate of the required frequency of an AC electrical signal for which the electrical double layer doesnt have time to form, and its effects can be ignored.

A microfluidic device, based on electrophoretic transport and electrostatic trapping of charged particles, has been developed for continuous capture and concentration of microorganisms from water. A generic design, utilizing mobility and zeta potential measurements of various microorganisms exposed to different environmental conditions and physiological states, was employed. Water and buffer samples at pH values ranging from 5.2–7.0 were seeded with bacteria (E. coli, Salmonella, and Pseudomonas) and viruses (MS-2 and Echovirus). Negative control and capture experiments were performed simultaneously using two identical devices. Both culture based methods and real-time PCR analysis were utilized to characterize the capture efficiency as a function of time, flowrate, and applied electric field. Based on differences between the capture and negative control data, capture efficiencies of 90% to 99% are reported for E. coli, Salmonella, Pseudomonas, and MS-2, while the capture efficiency for Echovirus was around 75%. Overall, the device exhibits 16.67 fold sample volume reduction within an hour at 6 mL/hr. This results in a concentration factor of 15 at 90% capture efficiency. Direct quantification of capture on the anode of the prototype microfluidic device was also performed by particle tracking using fluorescent microscopy. Based on image processing, the capture data at different locations on the electrode surface is quantified as a function of the wall shear stress at these locations, which is calculated using CFD simulations. Finally, the Faradaic processes in the microchannel due to electrochemical reactions are studied to predict the amount of electrophoresis in the system. Scaling of the device to sample 5 L/hr can be achieved by stacking 835 identical microchannels. Power and wetted volume for the prototype and scaled devices are presented. The device can thus function either as a filtration unit or as a sample concentrator to enable the application of real-time detection sensor technologies. The ability to continuously sample water without chemical additives facilitates the use of this device in drinking water distribution systems. This work constitutes the first step in our development of a continuous, microbial capture and concentration system from large volumes of potable water.

This thesis presents a novel micron-sized trapezoidal plasmonic waveguide design, called an Interface Plasmon Polariton waveguide. The guiding mechanism is explained using an effective index method and validated by simulations. The mode cut-off conditions and single-mode guiding properties are both determined using simulation and experimentally demonstrated. The waveguides have a long 1 mm propagation distance at 1550 nm wavelengths. Using this IPP waveguide, novel dielectric rib, dielectric varying-density hole-array, and metal-groove Bragg grating $\emph{in vitro}$ sensors are designed, fabricated, and characterized. The devices have a 1100 nm/RIU sensitivity and 0.006 RIU sensing resolution obtained from measurements and are validated by theory. The IPP sensors developed in this thesis not only offer competitive plasmonic sensitivity, sensing resolution, signal to noise ratio, result reproducibility, and reusability, they are also easy to fabricate and simple to package. Therefore, these new sensor designs are an enabler for lab-on-a-chip platforms to adapt plasmonic technology.

This thesis presents a novel micron-sized trapezoidal plasmonic waveguide design, called an Interface Plasmon Polariton waveguide. The guiding mechanism is explained using an effective index method and validated by simulations. The mode cut-off conditions and single-mode guiding properties are both determined using simulation and experimentally demonstrated. The waveguides have a long 1 mm propagation distance at 1550 nm wavelengths. Using this IPP waveguide, novel dielectric rib, dielectric varying-density hole-array, and metal-groove Bragg grating $\emph{in vitro}$ sensors are designed, fabricated, and characterized. The devices have a 1100 nm/RIU sensitivity and 0.006 RIU sensing resolution obtained from measurements and are validated by theory. The IPP sensors developed in this thesis not only offer competitive plasmonic sensitivity, sensing resolution, signal to noise ratio, result reproducibility, and reusability, they are also easy to fabricate and simple to package. Therefore, these new sensor designs are an enabler for lab-on-a-chip platforms to adapt plasmonic technology.

A microfluidic device, based on electrophoretic transport and electrostatic trapping of charged particles, has been developed for continuous capture and concentration of microorganisms from water. A generic design, utilizing mobility and zeta potential measurements of various microorganisms exposed to different environmental conditions and physiological states, was employed. Water and buffer samples at pH values ranging from 5.2–7.0 were seeded with bacteria (E. coli, Salmonella, and Pseudomonas) and viruses (MS-2 and Echovirus). Negative control and capture experiments were performed simultaneously using two identical devices. Both culture based methods and real-time PCR analysis were utilized to characterize the capture efficiency as a function of time, flowrate, and applied electric field. Based on differences between the capture and negative control data, capture efficiencies of 90% to 99% are reported for E. coli, Salmonella, Pseudomonas, and MS-2, while the capture efficiency for Echovirus was around 75%. Overall, the device exhibits 16.67 fold sample volume reduction within an hour at 6 mL/hr. This results in a concentration factor of 15 at 90% capture efficiency. Direct quantification of capture on the anode of the prototype microfluidic device was also performed by particle tracking using fluorescent microscopy. Based on image processing, the capture data at different locations on the electrode surface is quantified as a function of the wall shear stress at these locations, which is calculated using CFD simulations. Finally, the Faradaic processes in the microchannel due to electrochemical reactions are studied to predict the amount of electrophoresis in the system. Scaling of the device to sample 5 L/hr can be achieved by stacking 835 identical microchannels. Power and wetted volume for the prototype and scaled devices are presented. The device can thus function either as a filtration unit or as a sample concentrator to enable the application of real-time detection sensor technologies. The ability to continuously sample water without chemical additives facilitates the use of this device in drinking water distribution systems. This work constitutes the first step in our development of a continuous, microbial capture and concentration system from large volumes of potable water.

Scanning Probe Lithography (SPL) is a versatile nanofabrication platform that leverages microfluidic “ink” delivery systems with Scanning Probe Microscopy (SPM) for generating surface-patterned chemical functionality on the sub-100 nm length scale. One of the prolific SPL techniques is Dip Pen Nanolithography™ (DPN™). High resolution, multiplexed registration and parallel direct-write capabilities make DPN (and other SPL techniques) a power tool for applications that are envisioned in micro/nano-electronics, molecular electronics, catalysis, cryptography (brand protection), combinatorial synthesis (nano-materials discovery and characterization), biological recognition, genomics, and proteomics. One of the greatest challenges for the successful performance of the DPN process is the delivery of multiple inks to the scanning probe tips for nano-patterning. The purpose of the present work is to fabricate a microfluidic ink delivery device (called “Centiwell”) for DPN (and other SPL) applications. The device described in this study maximizes the number of chemical species (inks) for nanofabrication that can be patterned simultaneously by DPN to conform the industrial standards for fluid handling for biochemical assays (e.g., genomic and proteomic). Alternate applications of Centiwell are also feasible for the various envisioned applications of DPN (and other SPL techniques) that were listed above. The Centiwell consists of a two-dimensional array of 96 microwells that are bulk micromachined on a silicon substrate. A thermoelectric module is attached to the back side of the silicon substrate and is used to cool the silicon substrate to temperatures below the dew point. By reducing the temperature of the substrate to below the dew point, water droplets are condensed in the microwell array. Microbeads of a hygroscopic material (e.g., poly-ethylene glycol) are dispensed into the microwells to prevent evaporation of the condensed water. Furthermore, since poly-ethylene glycol (PEG) is water soluble, it forms a solution inside the microwells which is subsequently used as the ink for the DPN process. The delivery of the ink to the scanning probe tip is performed by dipping the tip (or multiple tips in an array) into the microwells containing the PEG solution. This thesis describes the various development steps for the Centiwell. These steps include the mask design, the bulk micromachining processes explored for the micro-fabrication of the microwell array, the thermal design calculations performed for the selection of the commercially available thermoelectric coolers, the techniques explored for the synthesis of the PEG microbeads, and the assembly of all the components for integration into a functional Centiwell. Finally, the successful implementation of the Centiwell for nanolithography of PEG solutions is also demonstrated.

The overall goal of this work was to develop and utilize microfluidic models for investigating bacterial chemotaxis and biofilm formation - phenotypes that play key roles in bacterial infections. Classical methods for investigating chemotaxis and biofilm formation have many limitations and drawbacks. These include being unsuitable for investigating the effect of chemorepellents, non-quantitative readouts, and not accounting for interaction between hydrodynamics and biofilm formation. The novel microfluidic model systems for chemotaxis and biofilm formation developed in this study addresses these drawbacks. Chemotaxis model system development was done in three stages. We first developed two static chemotaxis model systems - the two fluorophore chemotaxis agarose plug assay and the mu Plug assay - for rapidly determining the extent of chemotaxis in a qualitative manner. A key feature of these model systems was the incorporation of dead cells and differential labeling with green and red fluorescent proteins for partitioning the effects of movement due to fluid flow from chemotaxis. The static systems were used to rapidly screen a wide range of conditions for use in the flow-based mu Flow chemotaxis model system. The effect of four major variables - cell preparation method, gradient strength, flow rate in the device, and imaging position - that influence the chemotactic response in the mu Flow was characterized using the repellent taxis from Ni^2 gradients as the model chemoeffector. Using the mu Flow chemotaxis device, we investigated the chemotaxis of Escherichia coli RP437 to different signals that are present in the human gastrointestinal tract and are likely to be mediators of infection through their effect on chemotaxis. Our data show that the bacterial signal indole is a repellent, while the signals autoinducer-2 (AI-2) and isatin are attractants for E. coli RP437. However, cells exposed to a competing gradient of indole and either AI-2 or isatin, attracts E. coli. The ?Flow device was also used to refute a long-standing view on how the repellent Ni2 is sensed in E. coli. Our data show that only the Tar chemoreceptor is needed for sensing Ni^2 and the nickel binding protein, NikA, and the Ni^2 transport system proteins, NikB and NikC, are not required for repellent taxis from nickel. A microfluidic biofilm model was also developed in this study and used in conjunction with a mathematical model to investigate biofilm formation and quorum sensing in closed systems (where biofilm growth and hydrodynamics are interdependent). The mathematical model predictions were experimentally validated using Pseudomonas aeruginosa PA14 in a microfluidic biofilm system at various flow rates.