Phase Separation in Two-phase Microfluidic Heat Exchangers

Download or read book Phase Separation in Two-phase Microfluidic Heat Exchangers written by Milnes P. David and published by Stanford University. This book was released on 2011 with total page 151 pages. Available in PDF, EPUB and Kindle. Book excerpt: Two-phase microfluidic heat exchangers have the potential to meet the large heat dissipation demands of high power electronics and computing systems. Two-phase cooling systems face practical challenges brought on by the growth and advection of the vapor phase in the confined geometries, which lead to large pressure drops, increased thermal resistance and the formation of detrimental flow instabilities. One proposed solution to these issues is phase separation, whereby the vapor is locally separated from the two-phase flow through a porous hydrophobic membrane. This dissertation describes a series of studies conducted to develop an understanding of the factors that influence vapor separation and its impact on the hydraulic and thermal characteristics of two-phase heat exchangers. Flow phenomena are a critical component in developing this understanding of phase separation. High speed visualization of adiabatic and diabatic vaporizing flows was carried out in a single 124[Mu]m by 98[Mu]m copper microchannel with a 65[Mu]m thick, 220nm pore diameter hydrophobic PTFE membrane wall. During adiabatic air-water flow, wavy-stratified and stratified flow dominated lower liquid velocities, while plug and annular type flows dominated at the higher velocities. Analysis found that air removal could be improved by increasing the venting area, increasing the trans-membrane pressure or using thinner, high permeability membranes. Diabatic water-vapor experiments with mass flux velocities of 140 and 340 kg/s-m2 and exit qualities up to 20% found that stratified type flows dominate at lower mass fluxes while cyclical churn-annular flow became more prevalent at the higher mass-flux and quality. The observed flow regimes are hypothesized to play a significant role in determining the pressure drop and heat transfer coefficient during flow boiling. To study the impact of various geometric and membrane factors on the performance of a phase separating microchannel heat exchanger dissipating 100W of heat, a numerical model incorporating vapor separation and transport during two-phase flow boiling in a microchannel was developed. The impact of substrate thermal conductivity and thickness, membrane permeability and thickness, liquid channel density, liquid and vent channel diameter and vent-to-liquid channel diameter ratio was studied and compared for a standard non-venting heat exchanger, a vapor venting heat exchanger and a non-venting heat exchanger occupying the same increased volume as the venting heat exchanger. The numerical study found that the venting heat exchanger had improved pressure drop and device temperatures for all tested conditions when compared against a standard heat exchanger but only under very limited conditions when compared against the volumetrically equivalent non-venting heat exchanger. The study indicates that the best venting heat exchanger performance is achieved when the membrane conductance is of the same order or higher than that of the microchannel; this can be achieved through the use of thin high permeability membranes coupled with small hydraulic diameter microchannels. Finally, a study was conducted to explore the fabrication methods to build a vapor separating heat exchanger and to quantify the operating performance of multichannel silicon and copper phase separating devices. A copper parallel microchannel heat exchanger with nineteen 130[Mu]m square microchannels was built and tested at heat fluxes of up to 820 kW/m2 and water mass fluxes of between 102 and 420 kg/s-m2. Normalized pressure drop was improved by as much as 60% and average substrate temperature by a maximum of 4.4°C between the non-venting control and vapor venting device under similar operating conditions. Comparison between the experimental results and simulation predictions found higher than expected pressure drop improvements at higher mass fluxes and poorer heat transfer coefficients at the lowest mass flux. Based on the flow phenomena study these discrepancies are believed to be due to the mass flux and vapor quality dependent two-phase flow structures. The encouraging experimental and numerical results motivate further study into phase separation methods, materials and flow physics. The development of a high performance phase separating heat exchanger, with the thermal benefits of two-phase boiling flow and the hydraulic benefits of single-phase liquid flow, would strongly enable the adoption and application of two-phase heat exchangers to provide effective and efficient cooling for next generation high power computing systems.