Study of the Impact of the Airflow and Filling Ratio on the Thermal Performances of a Two-Phase Immersion Cooling Prototype

• Published in: 2022 21st IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (iTherm) pp. 1 - 6
• Main Authors: Ghaffari, Omidreza, Sayed, Chady Al, Vincent, Manuel, Grenier, Francis, Jasmin, Simon, Frechette, Luc, Sylvestre, Julien
• Format: Conference Proceeding
• Language: English
• Published: IEEE 31.05.2022
• Subjects: Electronics cooling; Fans; Immersion cooling; Liquids; Microprocessors; multi-scale porous coating; Power demand; Prototypes; Thermal resistance
• ISSN: 2694-2135
• DOI: 10.1109/iTherm54085.2022.9899558

Evacuating high heat fluxes for new generation microprocessors is a challenge in electronics cooling applications. We conducted a thermal performance study on a novel two-phase thermosyphon prototype to cool high-power microprocessors in a 4U form factor. The prototype was mounted on a heater assembly for the thermal performance tests. A heat spreader was attached on top of the square 2.54 cm by 2.54 cm heater, with the top side of the heat spreader being enhanced with a 500 micron thick multi-scale electroplated porous copper coating. The localized boiling and condensation occurred inside the prototype, using the Novec™ 7000 liquid from the 3M Corporation. The prototype was air-cooled (using a single fan), and different air flows were tested, from 102 m 3 /h to 237 m 3 /h. The effect of the filling ratio was studied between 9% and 70%. At a filling ratio of 9 % and an airflow of 237 m 3 /h, the maximum power achieved was (525 ± 7) W (a heat flux of 81± 1.2 W/cm 2 ), the heater wall temperature was 75°C, and no dry out was observed. At an airflow of 102 m 3 /h at the same power level, the heater wall temperature was less than 83°C, so the prototype could be used even with low airflow rates if the fan power consumption needs to be reduced in specific electronics cooling applications. The total thermal resistance included the contributions of the heater body, its thermal interface material layer, the evaporator (boilerplate with porous coating), the condenser, and the air-cooled heat exchanger. At the highest tested airflow rate, the minimum achieved heater to ambient air thermal resistance was (0.096 ± 0.0018)°C/W for a filling ratio of 9%. The dominant thermal resistance was from the heater to the liquid, representing between 50% and 70 % of the total resistance in all the tested scenarios at different filling ratios.