The exponential increase in demand for artificial intelligence computing power has made chip heat dissipation a primary bottleneck limiting computing performance. To address the challenge of high heat flux dissipation in chips, this study utilizes a vapor chamber that leverages Laplace pressure, generated by surface wettability differences, to drive the movement of the internal working fluid. The investigation centers on bionic fishbone-like and feather-like unit patterns, designing four gas-liquid circulation modes by combining different pattern arrays and considering both condensation and evaporation limits. Visual flow analysis was performed on vapor chambers with hydrophilic-hydrophobic array surface combinations under heat loads from 0 to 70 W. The self-driven mechanism of the patterned pathways was elucidated through observations of liquid film thickness and flow morphology. The results demonstrate that the sample 4 featuring a divergent pattern on the evaporation and multi-point pattern on the condensation exhibited no liquid column formation at 70 W, and its vaporization height was reduced by 15.5% compared to the sample with a divergent evaporation pattern and centralized condensation. It confirm the superior thermal performance of the model 4 configuration (thermal resistance: 0.211 degrees C/W) and underscore the importance of optimizing condensation in the heat transfer design of vapor chambers.
