Active Microfluidic Cooling of Integrated Circuits

The thermal management of high heat fluxes is a critical roadblock in the way of higher-performance microelectronics. The ongoing reduction in microtransistor size translates into heat fluxes comparable to those encountered in nuclear reactions and rocket nozzles, but under much severer temperature constraints. Although the average heat flux may remain in the vicinity of 100 to 300 W/cm, the peak heat flux at localized hot spots may approach or exceed 1 kW/cm. In order to handle these heat fluxes properly at lower operating temperatures, integrated cooling technologies that remove heat closer to the source are required. Microfluidic cooling in a heat sink built within or attached directly to the silicon chip is one route to address this problem. Three primary microfluidic technologies have the potential to accommodate this very large heat flux. These include microjet impingement, spray cooling, and microchannel heat sinks [1]. Microjet impingement relies on the use of a high-speed liquid jet that emerges from a nozzle to reduce thermal boundary-layer thicknesses and increase convection coefficients at the incident surface. Despite being one of the leading cooling technologies, microjet impingement has serious drawbacks in terms of fluid recovery for open systems and temperature uniformity. Therefore, intricate architectures using multiple microjets and specially designed outlet ports are needed. This leads to designs that are highly optimized for very specific thermal operating conditions and are therefore not robust to time and spatially varying thermal loads. Similar to microjet impingement, spray cooling relies on the impingement of liquid onto the heated surface. Spray cooling relies on a blanket of liquid droplets rather than a continuous liquid jet striking the heated source, providing better surface coverage and temperature uniformity. Furthermore, rather than relying on the convective transport capabilities of the thin liquid film formed when the droplets hit the surface, spray cooling is achieved through the evaporation of the aforementioned film. The latent heat associated with the phase change translates into lower surface-temperature requirements, another advantage over microjet impingement cooling. This makes spray cooling one of the leading contenders in the race to achieve heat-flux removal values exceeding 1 kW/cm. However, it has serious drawbacks in terms of the pressures required to achieve proper spray droplet breakup and the distribution needed to exploit the cooling potential of this technology fully.