Reliability of solder joints assessed by acoustic imaging during accelerated thermal cycling

A robust feature extraction and blob analysis system was developed for high quality solder joint assessment during accelerated thermal cycling (ATC). Thermal cycling test performed on flip chip type devices were inspected by Acoustic Micro Imaging (AMI). An automated image analysis system has been developed to provide enhanced images and facilitate the solder joint inspection process. The proposed system consists of three stages: solder joint detection, feature extraction and defect evaluation. The existence and location of solder joints in an ultrasound C-scan image are through detection by a gradient-based circular Hough Transform. Subsequently, useful features and parameters are extracted based on a region growing algorithm. The method is verified by inspecting test samples before, during and after accelerated thermal cycling. Informative features and analysis results are presented. The results clearly distinguish between normal and abnormal joints and show the cracks between the chip-to-bump interfaces after accelerated thermal cycling tests.