Automated visual inspection of particle defect in semiconductor packaging

In semiconductor production processes, controlling and inspecting contamination particle defects are extremely important because even a small particle within any stage of the process can remarkably affect the quality of the final products. Particle contamination can be critically detrimental in every process, thereby reducing production yield in semiconductor processes. In this study, we investigated the correlation between the actual defect rate and the probability of contamination particle defect observed by a commercially available automated visual inspection (AVI) system in semiconductor backend processes. During mass production, we observed that contamination particles produced in a thermal process were transported to various locations and caused defects. Particles sized 45 mu m were observed most frequently compared with the actual contamination particles and AVI images. To effectively detect particle defect on wafer surfaces, particles smaller than 100 mu m should also be considered. The hallmark of this study is that we effectively controlled particles larger than 50 mu m using our AVI equipment after the die attach approach to reduce defects in the wire bonding process in advance. We provide monitoring methods for contamination control of particles present in the thermal process on the AVI system applied in mass production processes. Finally, we suggest a plausible entrainment pathway of the contamination particles and present visual images of actual contamination particles observed using an optical microscope.