Data-center cooling represents 30–40% of the total data-center facility energy consumption, and the rapid growth of high-power processors is exceeding the practical limits of pumped single-phase liquid cooling. Capillary-driven two-phase cooling offers a path to chip-scale thermal management with orders-of-magnitude lower pumping power, but its area-scaling to practical chip areas remains constrained by the liquid transport limit of wick. In this study, a silicon pin array coated with a copper inverse opal (CIO) wick and copper wire-mesh (CWM) 3D manifold are combined to construct capillary-driven two-phase micro-coolers by decoupling liquid feeding and vapor extraction. We systematically investigated the limits of area scaling using three combinations of heated area and 3D manifold footprints of (case A: 5 × 5 mm 2 : 5 × 5 mm 2 ), (case B: 10 × 10 mm 2 : 10 × 10 mm 2 ), and (Case C: 5 × 5 mm 2 : 10 × 10 mm 2 ), respectively. For all cases, A, B, and C, we achieved critical heat fluxes (CHF) of ≈ 220 and ≈ 440 Wcm −2 at heated-area mass fluxes of 5 and 10 g/min/cm 2 , respectively, with full mass flow utilization for phase change. When excess fluid is supplied to the CWM 3D manifold, for cases A and C, slightly higher CHF level of ≈ 550 Wcm −2 is achieved. At this heat flux and mass flow rate, high-speed camera images reveal the formation of a liquid-vapor phase front, indicating excess liquid that is not fully vaporized. However, for case B (10 × 10 mm 2 : 10 × 10 mm 2 ), the CHF value is pinned at ≈ 440 Wcm −2 , which indicates that CWM 3D manifold operates near its capillary limit, as the viscous pressure drop is highly non-linear for higher mass flow rates along CWM for the larger footprint CWM 3D manifold. At the most efficient operating point, the coolers achieved two-phase heat transfer coefficients, HTC 2-ϕ ranging from 0.4 to 0.5 MWm −2 °C −1 for cases A and B. The findings in this study provide crucial geometric design criteria for optimizing the hydraulic performance of two-phase micro-coolers utilizing capillary-based 3D manifold and high-performance porous wick.
No takes yet. Share an insight, caveat, or question.
Kwon et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: