Control of refrigerant expansion offers improved system efficiency and reduced noise among other benefits. The cost of traditional active expansion devices, however, has limited their use to large HVAC&R systems. The design, fabrication and testing of a thermopneumatic microfabricated valve for controlling refrigerant mass flow rate during expansion is presented in this dissertation. This device was fabricated through nickel electroplating within very thick SU-8 molds, thereby realizing expansion devices useful for small HVAC&R applications through an inexpensive modified UV-LIGA process.

This work begins with process development to make meso-scale nickel electroplating within SU8 micromolds a feasible option. Then, two test apparatuses were constructed and used to benchmark the performance of a prototype; one to control the flow of compressed air, one to control refrigerant expansion. Next, three dimensional CFD simulations were performed on the flowfield within the device at various levels of actuation to predict the device's ability to control compressed air flow. A numerical code was also developed to predict the device's temporal response and relationship between actuation level and power input.

The assembled prototype was demonstrated on the air flow test bench. The prototype was able to reduce the mass flow rate of the compressed air by 22 % at the conditions used in the CFD analysis. The performance was then demonstrated in a 1.5-2 kW R134a vapor compression system. Both steady state and transient response were characterized. Steady state data showed that the mass flow rate of refrigerant could be effectively controlled using the valve. The level of refrigerant subcooling defined the magnitude of the response. Steady state data taken at 750 kPa inlet pressure shows the mass flow rate was reduced by 4.2 % at 1 ºC subcooling and up to 10.8 % at 5 ºC subcooling for a given level of actuation. Transient system response was characterized using cyclic actuation of the device in the HVAC system. The change in capacity was approximately 5 %, at the conditions used during these tests. Data from the transient response tests showed the device's time constant to be within 11 % of the value predicted in the simulations.