Strain-rate Dependent Fracture Load Prediction of Lead-free Solder Joints

The critical strain energy release rate at crack initiation, Jci, has been shown to govern the fracture of short solder joints. The present work measured Jci for 2 mm long SAC305 solder joints as a function of the strain rate and phase angle using DCB specimens. A drop tester was designed and built to perform the experiments at higher strain rates. Jci increased about 70% from quasi-static to strain rates of 0.05-1 s-1 (intermediate strain rates), and then decreased by more than 67% from intermediate strain rates to 42 s-1. In the all strain rate regimes, Jci substantially increased with the phase angle. Next, the effects of five significant processing parameters on Jci were investigated. Jci was found to be only a weak function of the solder composition, time above liquidus, and aging. The effects of solder thickness and local end geometry were found to be significant only at the higher strain rates. To investigate the discrepancy in fracture behavior of solder joints between DCB and Arcan specimens, Jci was measured using both of these specimens. Jci values obtained from Arcan-type specimens were much smaller than those obtained from DCB specimens. The modeling showed that the differences in the degree of constraint and plastic dissipation in the solder layers of these two specimens were responsible for this large difference. Finally, BGA test vehicles were assembled and prepared in different geometries. The specimens were loaded at two strain-rate regimes. A strain-rate dependent mode I cohesive zone model (CZM) was shown to be sufficient to predict the fracture load of these joints. Two different modeling approaches were used: the first method utilized mode I Jci obtained from the 2 mm model DCB specimens as the fracture energy in the CZM and the peak traction was calibrated using the calibration specimen; in the second approach, the peak traction was assumed to be the solder yield stress at the corresponding strain rate regime and the mode I fracture energy was obtained using the calibration specimen. The predicted fracture loads obtained from both methods were in reasonable agreement with the measured forces.

Department dc:contributor.department

Mechanical and Industrial Engineering

Year dc:date.issued

2016

Author dc:creator

• Nourani, Amir

Advisor dc:contributor.advisor

• Spelt, Jan K

Handle dc:identifier.uri

http://hdl.handle.net/1807/73112

OAI identifier oai:identifier

oai:utoronto.scholaris.ca:1807/73112