Detection of Void Density in the Through Silicon Via Using Artificial Neural Network

3-D integration of integrated circuits based on through silicon via (TSV) could achieve high-performance, multifunctional and heterogeneous microsystem. However, various defects inside TSV may degrade system performance, it is essential to detect these defects for system design and process improvement. In this paper, we propose a defect detection method using artificial neural network (ANN) to detect the volume density of voids in the TSV. Different ANNs are designed and trained, their performance is compared. The results demonstrate the detection method could predict void density accurately.