Lead-free solders are soon to become the standard for use in electronics due to legislation prohibiting the use of lead. Experts generally agree that various Sn-Ag-Cu alloys are the best alternatives to the formerly ubiquitous eutectic Sn-Pb solder. To facilitate adoption of the 95.5Sn-3.9Ag-0.6Cu solder, a research partnership was created between Sandia National Laboratories and Stanford University to develop a methodology for predicting fatigue behavior of this alloy under isothermal and thermal mechanical conditions.; To begin, Sandia researchers conducted extensive material testing on the 95.5Sn-3.9Ag-0.6Cu alloy including; compressive stress-strain and creep testing of as-cast and thermally-aged cylinders, isothermal fatigue testing of joints in a double lap-shear (DLS) test vehicle, and thermal mechanical fatigue (TMF) testing of joints in a ball grid array (BGA) package.; To facilitate computational modeling, a Sandia developed unified creep plasticity (UCP) constitutive model is implemented in the commercial finite element analysis (FEA) software ANSYSRTM. A continuum damage model is next coupled to the UCP model (the unified creep plasticity damage UCPD model) to capture stress softening during fatigue. Material and damage parameters are fit to capture temperature effects from -55°C to 160°C. The UCPD model (mathematical structure and 95.5Sn-3.9Ag-0.6Cu parameters) predicts the true physical response of lead-free solder as demonstrated by comparison of simulation and analogous experimental results.; Next, DLS fatigue crack length empirical data are interpreted and discussed. A new crack length prediction methodology is formulated by interpreting crack growth in terms of viscoplastic strain energy density increments and correlated to the empirical data. The prediction methodology is capable of predicting crack growth in isothermally fatigued 95.5Sn-3.9Ag-0.6Cu joints as illustrated by comparison with empirical data.; Finally, the prediction scheme is tested and proven capable of predicting crack growth for the more general case of TMF. First, TMF crack length data from a BGA package is analyzed and discussed. Next, FEA models are built and exercised to simulate TMF deformation in the BGA assembly and results are extracted to inform the prediction methodology. A comparison with available data verifies that the numerical scheme is capable of predicting fatigue crack lengths in 95.5Sn-3.9Ag-0.6Cu solder joints undergoing TMF.
展开▼
