Optimization of a pulsed carbon dioxide snow jet for cleaning CMOS image sensors by using the Taguchi method

This study demonstrated the optimization of a pulsed carbon dioxide (CO2) snow jet system for the removal of particles on the surface of complementary metal oxide semiconductor (CMOS) image sensors by using the Taguchi method. The parameters of the CO2 snow cleaning system, which can have an influence on the residue rate of particles, were optimized, resulting in optimal values of 15° incident angle, 40 mm cleaning distance, 0.30 mm orifice size, and a 50 ms time-base with above 95% confidence. A novel pulsed CO2 snow jet, triggered by ordered pulsing signals, was introduced to the experiments for solving the cryogenic effect on the surface. The pulsed CO2 snow jet gained a total of 10.6 °C difference in average temperature on the substrate surface. So, not only did we achieve less (even none at all) hot air consumption, but we also achieved less CO2 consumption. Due to the optimization, the average residual particle rate can be controlled, and all particles larger than 2 μm will be removed.