Thermal Fatigue Life Prediction Model of CCGA Tin-Lead and Lead-Free Interconnects

Abstract

A thermal fatigue life prediction model of a ceramic column grid array (CCGA) solder joint assembly has been developed when the 90Pb/10Sn solder columns of the CCGA package are soldered onto the printed circuit board with either tin-lead or lead-free solder paste. This model was evolved from an empirically derived formula by correlating the solder nonelastic strain energy density increment to the fatigue life test data. To develop the solder joint fatigue life prediction model, a nonlinear finite element analysis (FEA) was conducted using the ABAQUS computer code. A thermal fatigue life prediction model was then established. The test results, obtained from various sources in which tin-lead and lead-free solder pastes on PCB were used, combined with the FEA derived nonelastic strain energy density per temperature cycle, ΔW, were used to calibrate the proposed life prediction model. In the analysis, 3-D finite element global- and sub-modeling techniques were used to determine the ΔW of the CCGA solder joints when subjected to temperature cycling. The analysis results show that: 1) solder joint would typically fail across solder column instead of along solder pad interfaces; and 2) higher nonelastic strain energy densities of solder occur at the solder columns at the package corners and these solder joints would fail first. These analysis predictions are consistent with the test observations. In the model calibration process, the 625- and 1657-pin CCGA test results, which were cycled between 20°C/90°C, 0°C/100°C, -55°C/110°C, or -55°C/125°C, were reasonably well correlated to the predicted values of ΔW. Therefore, the developed life prediction model could be used and is recommended to serve as an effective tool to determine the integrity of the CCGA solder joints during temperature cycling. In addition, the following future work is recommended: 1) selecting more study cases with various solder joint configurations, package sizes, environmental profiles, etc. to further calibrate this life prediction model; 2) using this model to conduct parametric studies to identify critical factors impacting solder joint fatigue life and then seek an optimum design; and 3) developing a simplified method instead of the FEA approach to make preliminary thermal fatigue life estimates of the CCGA solder joints.