Hybrid FEM-ANN-PSO Method To Optimize The Structural Parameters Of Wafer-Level Chip Scale Package (WLCSP) For High Reliability

摘要

Fatigue life prediction of electronic devices is of great importance in both research and industry. One of the factors which have a powerful influence on the fatigue life of WLCSP package are the structure parameters of the WLCSP. This paper developed a hybrid method for fatigue life prediction of WLCSP under thermal cycling and optimization of the structure parameters of WLCSP for obtaining high reliability of WLCSP. The fatigue life of one 3D WLCSP model under thermal cycling was calculated by using finite element method (FEM). Then the fatigue life of different WLCSP structures under different chip thickness, PCB thickness and solder joint pitches were obtained by the FEM. Then Artificial neural network (ANN) model was trained and tested by the data, which can be employed to describe the relation between different structure parameters with the fatigue life of WLCSP well with the predict accuracy 89.2% of the testing data. The proposed ANN model combined with the Particle Swarm Optimization (PSO) algorithm was used to select optimal structure parameters under certain constrains for obtaining the longest fatigue life. The selected structure parameters were verified by theoretical analysis. The proposed integrated (FEM-ANN-PSO) approach was found efficient and robust to predict the fatigue life of WLCSP and optimize the structure parameters of WLCSP as the selected structure parameters were found to give the highly reliability of WLCSP.