Abstract
The increasing trend of power densities in high-performance computing, driven by artificial intelligence, machine learning, and cloud computing, necessitates advanced thermal management solutions to maintain operational stability and energy efficiency. This study examines the effectiveness of cooling a 1.5 U simulated copper microchannel chip compared to a plain chip. Both chip types were tested with and without configurations for dual taper microgaps to enhance the heat transfer performance of a boiling chamber (BC). Experimental investigation was conducted using 500 μm wide × 400 μm deep microchannels separated by 200 μm fins. Varying inlet gaps (0.5-4 mm) and taper lengths (8.25 mm and 16.5 mm) with a taper angle of 3 deg were employed in dual taper configuration. Their impact on critical heat flux (CHF) and subcooled boiling dynamics was investigated. Microchannels provided considerable performance enhancement over a plain surface with or without the dual taper microgap. The findings demonstrate that smaller inlet gaps (0.5-1 mm) and longer taper lengths (16.5 mm, with central liquid inlet) significantly enhance nucleate boiling. These configurations improve vapor escape and delay CHF through subcooled boiling and submerged condensation. However, a lower CHF was noted due to vapor agglomeration within the microgap. The 80% fill ratio microchannel chip exhibited the highest CHF as subcooled boiling increased liquid replenishment and prevented vapor stagnation. Similarly, lower coolant temperatures (20-30 °C) enhanced boiling performance, where submerged condensation accelerated bubble collapse and improved heat dissipation efficiency in lower surface temperatures.