Date of Award

Winter 1997

Document Type


Degree Name

Doctor of Engineering (DEng)


Micro and Nanoscale Systems

First Advisor

Randall Barron


The purpose of this study was to investigate fluid mechanic and heat transfer characteristics of two-phase two-component flow in rectangular microchannels. Experiments were conducted using rectangular aluminum channels with hydraulic diameters ranging between 56 $\mu$m and 256 $\mu$m and aspect ratios which varied from 0.5 to 1.5. Both single- and two-phase tests were conducted using water and gaseous argon, helium, and nitrogen as the working fluids. The Reynolds number for both types of experiments ranged from approximately 50 to nearly 10,000. The Nusselt number ranged between 0.0002 and 70. The single- and two-phase experimental data were empirically correlated, using parameters derived from a dimensional analysis. Experimental data was also used to correlate the unknown variables in derived analytical expressions.

Both single- and two-phase tests yielded excellent correlations of the friction factor. For Nusselt number, the correlations were fair to poor. Reynolds number and the combination of Reynolds number and Prandtl number were the dominant parameters in the prediction of pressure drop and heat transfer rate, respectively, in both single- and two-phase flows. The pressure drop predictions based on the semi-empirical relations by Martinelli for two-phase flows were shown to substantially overpredict the pressure drop measured in these experiments.

Other findings showed that for single-phase flow, the transition from laminar to turbulent regimes of the friction factor was suppressed as the channel hydraulic diameter decreased. For gases, the suppression occurred in varying degrees between hydraulic diameters of 80 $\mu$m and 150 $\mu$m, with no transition seen below 80 $\mu$m. For water, no transition was seen for any of the channel configurations tested. Two-phase friction factor data showed a definite transition from laminar to turbulent regimes at a Reynolds number of 3,000 for all channel configurations tested. It is believed that the transition was due to the intense pressure fluctuations associated with two-phase flows. For the Nusselt number, both single- and two-phase data was seen to parallel, though lower, the macroscale turbulent regime predictions. Also, no transition from laminar to turbulent regimes was seen in the Nusselt number data over the range of Reynolds numbers tested.