Numerical simulation on heat pipe for high power LED multi-chip module packaging

Abstract

Light emitting diode (LED) as the new light source has the advantages of power saving, environment-friendly, long lifetime and no pollution compared with fluorescent and incandescent lights. But the disadvantage of LED is low light lumen that only 10%~20% input power transform into the light, and 80%~90% into the heat. The junction temperature of LED is so high as to induce the lifetime declining rapidly, luminous decay and reliability decreasing. Therefore, the effective thermal management is very important for the LED light system. In this work, a new packaging architecture the system in package (SiP) configuration is used in the high power LED packaging. The light system consists of nine chips that each chip is 1.2W. Copper/water miniature heat pipe (mHP) is chosen to dissipate heat based on the LED packaging structure and the input power of the system. The principles of the heat pipe are investigated to design and select the structure and size of the heat pipe. Capillary limit and boiling limit of the heat pipe are calculated to determine the maximum heat transfer and verify the design of the heat pipe. The heat pipe is seen as the thermal superconductor in axial, which take the place of the process of the phase exchange in the pipe. The axial thermal resistance of mHP estimated by the net of the thermal resistance is 0.15°C/W approximately. The system level heat and temperature distribution are investigated using numerical heat flow models. In this analysis, 3D finite volume model is developed to predict the system temperature with Icepak which is the professional software to analyze the temperature field of electronics. The result shows that the junction temperature of the source is under 70°C at the natural convection which is satisfied with the requirement of the LED working at under 120°C. It shows that the heat pipe is the effective solution for the LED light application dissipation. For the lower junction temperature, three factors including the height, the thickness and the fin numbers of the heat sink, respectively, are considered to be optimized by DOE (design of experiment). With the simulation results of Icepak, the optimal scheme that the lower junction temperature is 56.7°C obtained by the combination of optimization levels.