The behavior of lead-free solder alloys under complex loading scenarios is still not well understood. Common damage accumulation rules fail to account for strong effects of variations in cycling amplitude, and random vibration test results cannot be interpreted in terms of performance under realistic service conditions. This is a result of the effects of cycling parameters on materials properties. These effects are not yet fully understood or quantitatively predictable, preventing modeling based on parameters such as strain, work, or entropy. Depending on the actual spectrum of amplitudes, Miner’s rule of linear damage accumulation has been shown to overestimate life by more than an order of magnitude, and greater errors are predicted for other combinations. Consequences may be particularly critical for so-called environmental stress screening. Damage accumulation has, however, been shown to scale with the inelastic work done, even if amplitudes vary. This and the observation of effects of loading history on subsequent work per cycle provide for a modified damage accumulation rule which allows for the prediction of life. Individual joints of four different Sn-Ag-Cu-based solder alloys (SAC305, SAC105, SAC-Ni, and SACXplus) were cycled in shear at room temperature, alternating between two different amplitudes while monitoring the evolution of the effective stiffness and work per cycle. This helped elucidate general trends and behaviors that are expected to occur in vibrations of microelectronics assemblies. Deviations from Miner’s rule varied systematically with the combination of amplitudes, the sequences of cycles, and the strain rates in each. The severity of deviations also varied systematically with Ag content in the solder, but major effects were observed for all the alloys. A systematic analysis was conducted to assess whether scenarios might exist in which the more fatigue-resistant high-Ag alloys would fail sooner than the lower-Ag ones.
Read full abstract