Experimental investigation of the thermal resistance of advanced two-phase thermosyphon heatsinks

• Published in: Applied thermal engineering 2024-02, Vol.238, p.122108, Article 122108
• Main Authors: Vincent, Manuel, Ghaffari, Omidreza, Nabavi Larimi, Yaser, Grenier, Francis, Jasmin, Simon, Fréchette, Luc, Sylvestre, Julien
• Format: Article
• Language: English
• Subjects: Dielectric liquid; Electronics cooling; Engineering Sciences; Heat transfer characteristics; Thermal resistance; Thermosyphon
• ISSN: 1359-4311; 1873-5606
• DOI: 10.1016/j.applthermaleng.2023.122108

•Comparison of four variants of a new air-cooled reflux thermosyphon design.•The studied design parameters were the condenser volume and boiling area.•A thermal resistance of 0.064 K/W was achieved at 128 m3/h airflow under 500 W power.•A formula predicting the thermal resistance was fitted to the experimental results.•The model has a maximum relative deviation of 10% over the experimental results. In this study, the heater-to-air thermal resistance of an air-cooled reflux thermosyphon with internal pool boiling and confined condensation was studied relative to four parameters the heater input power, the fan input power, the integrated heat spreader area and the condenser volume. Four different prototypes were identically designed except for their integrated heat spreader area and condenser volume. The prototypes were tested for a heater input power ranging from 200 W to 500 W and a fan input power from 2.5 W to 18 W. Each prototype was filled at its optimal filling ratio with a dielectric liquid that was degassed before the measurements. The heater-to-air thermal resistance of the prototypes was broken down into three intermediate thermal resistances: the heater-to-liquid, liquid-to-condenser and condenser-to-air thermal resistances A model was constructed to predict each intermediate thermal resistance. The parameters of this model were fitted to the experimental data, to achieve a mean relative deviation and a maximum relative deviation of 3.4 % and 10 % respectively when comparing the model to the experimental data.