Microstructure and Thermal Fatigue Life of BGAs with Eutectic Sn-Ag-Cu Balls Assembled at 210~C with Eutectic Sn-Pb Solder Paste

Abstract

Conventionally, it has been understood that in order to assemble BGAs with eutectic Sn-Ag-Cu solder balls in eutectic Sn-Pb solder paste, a reflow profile having a peak temperature higher than 217/spl deg/C, which is the melting point of eutectic Sn-Ag-Cu, would be required to achieve complete mixing of the Sn-Pb paste and the Sn-Ag-Cu ball. Nevertheless, it was observed in this study that solder joints with uniform microstructure may be achieved with a peak reflow temperature of about 210/spl deg/C. For further understanding of this phenomenon, DSC testing was performed on eutectic Sn-Ag-Cu spheres (of 0.406-0.76 mm diameters) placed over eutectic Sn-Pb solder paste of different volumes. It has been observed that the mixing of the Sn-Pb and Sn-Ag-Cu was accomplished in about 15 to 25 sees at a peak temperature of 210/spl deg/C. Thermodynamic calculations published in the literature indicate that the liquidus temperature of completely mixed Sn-Pb and Sn-Ag-Cu is above 210/spl deg/C. However, above 205/spl deg/C the volume ratio of solid phase to liquid phase is very small, or practically negligible. The metallograph of Sn-Pb/Sn-Ag-Cu solder joints reflowed at 210/spl deg/C peak temperature shows uniform microstructure. However, the distribution of Pb-rich phase inside the alpha Sn for the 210/spl deg/C reflowed Sn-Pb/Sn-Ag-Cu solder joints was somewhat different from those of the 240/spl deg/C reflowed Sn-Pb/Sn-Ag-Cu solder joints. Accelerated thermal cycling tests (0-100/spl deg/C and -40-125/spl deg/C) have been conducted on CSPs with eutectic Sn-Ag-Cu balls assembled on PCBs (of 0.093" and 0.125" thick) with eutectic Sn-Pb solder paste. SnPb and Pb-free control samples were also included. These parts were assembled in both single and double-sided board configurations and were reflowed with peak temperatures of 210/spl deg/C and 227/spl deg/ C. The ATC test results showed that the fatigue life of CSP assemblies reflowed at 210/spl deg/ C and 227/spl deg/C had no significant difference.