Fluid flow and heat transfer behavior of a liquid based MEMS heat sink having wavy microchannels integrating circular pin-fins

Abstract

MEMS heat sink having wavy microchannels integrating circular pin-fins for cooling of microelectronic chips is proposed and analyzed in this work; the pin-fins are placed at the valleys and peaks of the wavy structure. The pumping power and thermal resistance of the proposed MEMS heat sink is higher and lower than that of MEMS heat sink having straight microchannels. At Reynolds number of 250, the maximum temperature of the microelectronic chip, generating 106 W/m2, when using the proposed MEMS heat sink is only 0.77 of that when using MEMS heat sink having straight microchannels; this ratio lowers to 0.6 at Reynolds number of 1250. The thermal resistance of the proposed MEMS heat sink is only 0.72 and 0.43 of that of the MEMS heat sink having straight microchannels at Reynolds numbers of 250 and 1250, respectively. Regarding the pumping power of the proposed MEMS heat sink, it is higher than that of the MEMS heat sink having straight microchannels by 2 and 3.7 at the lowest and highest Reynolds numbers, respectively. This work also analyzes the contribution of geometric features on the behavior of the proposed MEMS heat sink in terms of maximum temperature of the microelectronic chip, thermal resistance, and pumping power as well as that of microchannel employed in the proposed MEMS heat sink in terms of Poiseuille and Nusselt numbers. Maximum temperature of the microelectronic chip and thermal resistance decrease along with increase in pumping power with increase in amplitude, diameter of pin-fins, and hydraulic diameter of microchannel as well as with decrease in wavelength. Poiseuille and Nusselt numbers increased with rise in amplitude, diameter of pin-fins, and hydraulic diameter as well as decrease in wavelength. In addition, experiments have been conducted on two designs of the proposed MEMS heat sink to generate data for purposes of validating the model. The experimental data are compared with data from simulations and the difference between the two are within the uncertainty of experimental data thereby validating the model.