Graduation date: 2008It is hypothesized that increases in pressure drop due to vapor generation during boiling in microchannels can be reduced by extraction of vapor at its point of inception. Ultimately, this local vapor extraction decreases the pressure drop required to drive the flow through a fractal-like branching, microchannel flow network within the heat sink. Indeed, by lowering the overall flow rate by vapor extraction the pressure drop can, in principle, be lowered to that of single-phase flow. In the current study, the feasibility of vapor extraction and its influence on the pressure drop across the microchannels are investigated. The concept also has the potential to separate flow independent of heat sink orientation or the influence of gravity. The fractal-like flow network used here is one that has been previously shown to reduce pressure drop and yield a more uniform surface temperature distribution for single-phase flows than that observed in parallel microchannel flow networks. The disk shaped heat sink was covered with two porous Nylon membranes with an average pore size of 0.45 microns that were backed with a porous aluminum block. Deionized, degassed water was used as the working fluid. A theoretical model was developed to predict the pressure drop across the flow network as a function of inlet mass flow rate, heat input, and pressure difference driving extraction across the membranes. Results of the model predictions are presented and discussed in both dimensional and non-dimensional format. Model predictions were used to discuss the trends and physical implications of local vapor extraction. It was observed that conditions existed where further increases in the membrane pressure difference had no influence vapor extraction and network pressure drop values. Experimental data was also collected and analyzed in order to assess the validity of the theoretical model. The predictive model and experimental data indicated that the two-phase network pressure drop is reduced by locally extracting vapor from the heat sink. The network pressure drop was reduced by as much as 80% theoretically and 50% experimentally in the predictive model and experimental data, respectively, for similar test conditions. A sensitivity analysis was performed on several parameters used in the model in an attempt to help explain some of the discrepancies observed between the experimental results and model predictions. Suggestions are made to improve the theoretical model pressure predictions and to investigate the influence of vapor extraction on the heat transfer coefficient values.