Experimentation and Multiphysical Modeling of Bioanalytical Microdevices

Bioanalytics involves quantitative measurements of complex biological samples that contain metabolites, DNA, RNA, and proteins. Efficient sample preparation for downstream analysis and sensitive detection of analytes can be achieved via bioanalytical microdevices. Fully realizing the potential of these devices requires tool characterization and bioprocess optimization, in addition to understanding device physics. Therefore, this thesis introduces multiphysical modeling and experimentation of microdevices, with applications to diabetes care and single-cell analysis. To understand the physics of viscometric glucose microsensors, this thesis presents a model of the sensor, which couples the fluid flow with vibrating diaphragms. The model is used to predict the sensor response to glucose via theory of squeeze-film damping and vibrations of pre-stressed plate. A first-principle-based model resulting from the theory can be evaluated from the device's geometric and material properties, and quantitatively determines the device response to vibrational excitations at varying glucose concentrations. Next, this thesis introduces a theoretical model for viscometric glucose microsensors that employ harmonic microcantilever oscillation in the sensing liquid. The presented model associates the unsteady Stokes equation with the motion of a bounded viscous liquid to understand the hydrodynamic impact on the cantilever. With a proper consideration of the viscosity and bounded geometry of liquid media, the model relaxes the thin-film assumption required for the diaphragm-based model, enabling an accurate representation of fluid-structure interactions based on fundamental structural vibration and fluid flow equations. Next, this thesis presents an experimental exploration of a hydrogel-based affinity microsensor for glucose monitoring via dielectric measurements. The microsensor incorporates a synthetic hydrogel that is attached to the device surface via in situ polymerization, which eliminates mechanical moving parts required in the viscometric glucose sensors. Changes in the dielectric properties of the hydrogel when binding reversibly with glucose molecules have been measured using a MEMS capacitive transducer to determine the glucose concentration. Experimental results demonstrate that in a glucose concentration range of 0–500 mg/dL and with a resolution of 0.35 mg/dL or better, the microsensor exhibits a repeatable and reversible response, and can potentially be useful for continuous glucose monitoring in diabetes care. Additionally, this thesis presents a microfluidic preprocessing method that integrates single-cell picking, lysing, reverse transcription and digital polymerase chain reaction to enable the isolation, tracking and gene expression analysis at single-cell level for individual cells. The approach utilizes a photocleavable bead-based microfluidic device to synthesize and deliver stable complementary DNA for downstream gene expression analysis, thereby allowing chip-based integration of multiple reactions and facilitating the minimization of sample loss or contamination. Finally, this thesis ends with concluding remarks and directions of future work towards continuous glucose monitoring and high-throughput single-cell genetic analysis.