AUTOMOTIVE PBGA ASSEMBLY AND BOARD-LEVEL RELIABILITY WITH LEAD-FREE VERSUS LEAD-TIN INTERCONNECT
Authors: Andrew Mawer Company: Motorola Semiconductor Date Published: 1/1/2001
Abstract: Many governments, particularly in Japan and Europe, have proposed laws reducing or eliminating the use of lead and other toxic substances in products in an effort to decrease landfill pollution and ground water contamination. Thus, there exists a need for lead-free components in order to comply with government standards and to meet market demands for “green” products. As the semiconductor industry moves towards environmentally friendly components, it is important that their assembly and reliability be well characterized. This paper will describe experiments performed to determine the interconnect reliability of a 1.0 mm pitch, four perimeter row, 324 PBGA for automotive applications with lead-free (tin-silver) versus conventional (tin-lead-silver) solder spheres. Additionally, assembly was carried out with lead-free versus eutectic solder paste for both sphere types. Two different PBGA substrate manufacturers were evaluated and the test boards used had an immersion nickel/electroless gold surface finish. The assembled boards were thermal cycled in both –40 to 125 and –50 to 150 degrees C until failure to compare lifetime of the four sphere and paste combinations. Solder joint failures were detected with continuous in-situ electrical monitoring and then verified with dye penetrant. Additionally, solder ball shear testing versus 125 and 150 degrees C bake out to 1,008 hrs was performed on lead bearing and lead-free solder balls. Ball shear as a function of number of reflows was also performed. Elemental analysis, including of intermetallics, was performed on solder joints as assembled to boards and after component aging and board-level thermal cycling. The data from the aforementioned experiments suggests that the lead-free 324 PBGA packages are at least as reliable as the lead containing ones.