As artificial intelligence grows, electronic devices are constantly being upgraded to improve performance to meet the surging demand in data computation. Efficient thermal management is critical to ensure devices operate stably. Manifold microchannels are used extensively as an excellent method of high heat flux cooling with low pumping power consumption. In this study, porous microchannel heat sinks with different manifold covers, named MMC-1 (50 % covering area) and MMC-2 (80 % covering area), were proposed for large-area heat flux removal to examine the influence of manifold covers on thermal performance and pumping power consumption. The inner wall surface of the manifold microchannel heat sinks was sintered with 1 μm copper powder to promote the occurrence of coolant nucleate boiling and enhance turbulent heat transfer. Water, HFE-7100, and a non-azeotropic immiscible binary mixture (HFE-7100/water) were compared as coolant in the manifold microchannel heat sinks, respectively. The flow rates were ranged from 50 to 150 L/h, and the total heat loads ranged from 250 to 2000 W over 5 cm heating area. The results showed that the thermal performance of MMC-2 was superior to that of MMC-1, which suggested that the larger manifold covering area will enhance heat transfer but also increase pressure loss. The MMC-2 heat sink can dissipating heat load up to 1700 W from 5 cm heating area (corresponding to = 360 W/cm) while maintaining the chip junction temperature below 85 °C with an extremely low pumping power below 0.32 W, which indicating a as high as 5695.

中文翻译：

---

以极低的泵浦功耗实现极低的热阻，以实现芯片尺寸尺度上的超高热通量去除

随着人工智能的发展，电子设备不断升级以提高性能，以满足不断增长的数据计算需求。高效的热管理对于确保设备稳定运行至关重要。歧管微通道作为一种低泵浦功耗的高热通量冷却的优异方法而被广泛使用。在本研究中，提出了具有不同歧管盖的多孔微通道散热器，称为MMC-1（50％覆盖面积）和MMC-2（80％覆盖面积），用于大面积热通量去除，以检查歧管盖的影响热性能和泵浦功率消耗。歧管微通道散热器的内壁表面烧结有1μm铜粉，以促进冷却剂核沸腾的发生并增强湍流传热。水、HFE-7100 和非共沸不混溶二元混合物（HFE-7100/水）分别作为歧管微通道散热器中的冷却剂进行比较。流量范围为 50 至 150 L/h，5 cm 加热面积上的总热负荷范围为 250 至 2000 W。结果表明，MMC-2的热性能优于MMC-1，这表明较大的歧管覆盖面积会增强传热，但也会增加压力损失。 MMC-2 散热器可以从 5 cm 加热区域消散高达 1700 W 的热负载（相当于 = 360 W/cm），同时以低于 0.32 W 的极低泵浦功率将芯片结温保持在 85 °C 以下，这表明高达5695。