Porous Rib Structure Optimization in Ultra-Thin Vapor Chambers

Sammanfattning

To manage increasing chip power densities in high-performance computing, vapor chambers have emerged as a leading passive cooling solution, but ultra-thin designs require support columns that obstruct vapor flow and compromise thermal uniformity. This thesis addresses how the number and arrangement of support ribs affect the coupled thermal-hydraulic behaviour. The primary goal was to develop a validated three-dimensional transient numerical model and use it to investigate the effect of rib number across multiple heat flux levels. The study is framed as a preliminary exploration of the design space rather than a formal optimization procedure, focusing specifically on the effect of rib number and arrangement while keeping rib thickness, cross-sectional shape, and spacing fixed. First, a no-rib baseline was simulated and confirmed to lose two-phase operation: lacking the parallel conduction paths and the auxiliary capillary return that the ribs provide, the evaporator superheats to roughly 197 °C and boils off the local liquid, yet the vapor never circulates or condenses back, so the chamber dries out. This verifies the necessity of support structures and validates the model's ability to capture this failure mode. Subsequently, a 12-rib, a 14-rib, and a symmetric 16-rib (4×4 matrix) configuration — the 12- and 14-rib layouts obtained by selectively removing non-critical ribs from the 16-rib baseline — were compared at heat fluxes of 80, 400, and 800 kW/m2. A key finding is that increasing the rib count does lower the peak temperature — at 400 kW/m2 the peak wick temperature falls from approximately 164 °C for the 12-rib layout to 94 °C for 14 ribs and 44 °C for 16 ribs — but the vapor volume fraction reveals that this reduction is driven by a change of operating regime rather than by improved two-phase performance. At the same heat flux the vapor volume fraction drops from full saturation (VF ≈ 1.0, VC-12) through partial saturation (VF ≈ 0.88, VC-14) to essentially zero (VF ≈ 0.005, VC-16): the denser the rib array, the more the chamber reverts from phase-change-dominated operation towards pure solid conduction. The low temperature of the 16-rib layout therefore reflects a near-solid metallic spreader rather than an efficiently working vapor chamber. Selecting the rib count is thus a trade-off between mechanical support and preserving an active vapor phase at the design heat flux. For the 12-rib configuration, which remains an active vapor chamber across the tested range, a tentative operating window of approximately 200–600 kW/m2 is suggested on the basis of the observed trends.