| Although studies of microscale flow have covered an extensive range of flow phenomena and flow conditions, one area that has received limited attention is bounded gas flows. This dissertation describes the applicability and characterizes the behavior of gas flows in microfluidic systems. We introduce the design of a pressure regulator based on thermally modulating the viscous energy dissipation of Venturi flow through a micronozzle. This component, which contains no moving parts, outputs both positive and negative pressures on the order of +/-1 kPa from a fixed positive pressure input of 110 kPa (gauge), using temperature control between 30 °C to 95 °C. To show the applicability of the Venturi microregulator for electronic flow control in gas-liquid microdevices, we demonstrate droplet manipulation in unbranched microchannels at linear velocities of 1 cm s-1 with a response time of 0.8 s. We then go on to develop a modular implementation for distributed pressure control in microfluidic networks. The pressure control module consists of four arrayed, individually addressable microregulators connected to a single external pressure input. We discuss the primary design considerations and possible performance optimizations for this device. As one example of possible applications to lab-on-a-chip platforms, we use the microregulator array for simultaneous manipulation of multiple droplets in a valveless, open microchannel network, highlighting key operations such as droplet merging, splitting, and transport. Finally, we characterize flow behavior in planar micronozzles at Reynolds numbers of ∼102--10 3 to better explain the mechanism behind the operation of the Venturi microregulator and, in a broader context, to advance fundamental understanding of compressible flow in microscale systems. This portion of our study compares simulation results and experimental measurements to established isentropic flow theory, providing a qualitative description of the deviation resulting from viscous wall effects. Changes caused by heat addition are also compared to Rayleigh flow expectations. |
