A study of the effect of exit boundary conditions on the performance of a spray cooling system

Spray cooling has remained one of the highest heat flux removal schemes across various applications such as cooling power electronics, lasers, high power LEDs, servers, supercomputers etc. Researchers have been researching extensively to obtain the maximum heat transfer coefficient. The effects of different parameters of the spray cooling system have been studied in the past. The focus of the present work is on the effect of exit boundary conditions on the performance of a spray cooling system, in particular, the method of draining the used coolant from the system. An 11 mm × 11 mm thermal test die was used as the simulated IC package to be cooled. The test die was mounted on to a 120 mm × 120 mm printed circuit board (PCB) by a flip chip attach. A hollow cone atomizer with central satellites, and orifice diameter of 0.258mm was used to spray water to cool the test die. There were two types of PCB boards used in the study, one of them was plain, wherein the used water had to travel until the edge of the PCB to be drained away, and the second type was a board, which had four petal-shaped grooves very near to the chip to enable quicker drainage of water. This latter design led to a thinner water film formed on the PCB. Experiments were carried out at 2, 4 and 6 bar inlet pressures for heat fluxes of 11, 22 and 33 W/cm². The steady-state temperature distribution was measured using 121 integral thermal diodes with a custom-built data acquisition board. It was observed that the nozzle spray pattern was different at 2, 4 and 6 bar. At 2 bar and 4 bar pressures, the temperature of the chip in the grooved board was lower compared to plain board as the film thickness formed on the chip was reduced and thus there was enhanced heat transfer. However, at 6 bar since the central part of the chip had a much smaller density of droplets, the thicker film formed on the plain board aided in cooling the chip and resulted in a lower temperature. Thus, it is concluded that the exit boundary conditions play an important role in the performance of spray cooling systems. Also, while using or comparing the existing heat transfer correlations it is important to make sure that the drainage method/conditions are the same. It is also proposed that to avoid confusion in future, a standardization of the test section is required. Alternatively, the film thickness should be reported along with other results for the sake of uniformity.