Amir Nourani and Jan K. Spelt
University of Toronto
Toronto, ON, Canada
The fracture of SAC305 solder was investigated as a function of strain rate using Cu-solder-Cu double cantilever beam (DCB) specimens joined with a series of 2 mm long discrete solder joints of 150 m thickness. The joints were then fractured with various strain rates under mode I and mixed-mode loading conditions.
The failure of each joint in the DCB was accompanied by a sharp drop in the applied load. These maximum loads were used to calculate the initiation critical strain energy release rate, Gci, of the solder joints using a finite element model. The results showed a substantial increase of about 75% in the solder joint fracture toughness at strain rates of 0.05 to 1 s-1 compared to that under quasi-static loading conditions.
This trend suggests that the solder Gci increases rapidly with strain rate whereas the Gci of the IMC is relatively independent of strain rate. Negligible changes in Gci were measured when the solder joint thickness was increased to 400 m. The dependence of crack initiation on the local shape of the solder joint at its end was investigated by fabricating the 2 mm long joints with either a square end (using Kapton tape) or a rounded end (using an embedded wire).
Interestingly, these two local geometries produced almost identical values of Gci, suggesting that initiation was not a strong function of the shape of the solder joint.
Variation in strain-rate can substantially change the initiation fracture toughness of lead-free solder joints. Experiments on SAC305 solder joints showed that the critical crack initiation strain energy release rate, Gci, increased by almost 75% when the loading rate was elevated from quasi-static (10-5 s-1) to intermediate strain rates (0.05 to 1 s-1).
Variations in solder joint thickness and the local end geometry of the joints were found to have a negligible effect on Gci at these strain rates. This is encouraging, since it should simplify fracture load predictions for actual solder joints in BGA and other components. For longer joints, however, R-curve toughening may cause a significant dependence of fracture toughness on solder thickness.
Being an electronic manufacturing manager and a highly qualified electronics tech. To me it seems as if lead free solders have their own short comings that are becoming evident with time. I also notice a higher end user failure rate indicating a shorter usable life of electronic devices that use certain lead free solders. I have seen many more cracked joints, TIN whiskers, cold solder joints, insufficient solder on joints as compared to similar devices using leaded solder. Just an observation over the last 10 years. (note when I repair I found that some devices work longer and are more durable after I use leaded solder, especially when I go over most all other leadfree connections with leaded.
