Creep Constitutive Models Suitable for Solder Alloys in Electronic Assemblies

Most solders used in electronic systems have low-melting temperature and hence experience significant amount of creep deformation throughout their life-cycle because typical operational and test conditions represent high homologous temperature. Phenomenological and mechanistic models used in the literature for predicting creep response of both bulk and grain scale specimens are reviewed in this paper. The phenomenological models reviewed in this paper are based on purely empirical observations of the creep deformation behavior or derived from qualitative interpretation of the underlying microscale mechanisms. These models have some intrinsic disadvantages since they do not have explicit mechanistic dependence on microstructural features. Therefore, the constitutive relations derived using the above models are difficult to extrapolate beyond the test conditions. This paper also reviews how some of the above limitations can be mitigated by using mechanistic or microstructurally motivated models. Mechanistic models are capable of estimating the material creep response based on the detailed physics of the underlying mechanisms and microstructure. The microstructure and constitutive response of the most popular family of lead-free solders, namely, SnAgCu (SAC) solders, are significantly different from those of previously used eutectic Sn37Pb solder. The creep deformation in Sn37Pb solder occurs primarily through diffusion-assisted grain-boundary sliding. In SAC solder joints, dislocation-based creep deformation mechanisms such as glide, climb, detachment, and cross-slip appear to be the dominant mechanisms in coarse-grained joints. Mechanistic creep models are therefore based on the deformation mechanisms listed above.

Properties of Two Lead-Free Solder Alloys and Comparison with Sn37Pb

The microstructures and properties of Sn3Ag2.8Cu and Sn3Ag2.8Cu-0.1Ce solder alloys were investigated by means of OM, SEM and EDX and compared to that of Sn37Pb. The results show that the wettability of Sn3Ag2.8Cu-0.1Ce is more favorable, Sn3Ag2.8Cu exhibits poorer wetting behaviour compared to that of Sn37Pb solder; the conductivities of Sn3Ag2.8Cu-0.1Ce and Sn3Ag2.8Cu soldesr are almost 20 percent and 8 percent higher than that of Sn37Pb respectively; the fractography of tensile specimen of Sn3Ag2.8Cu is smooth and light, and is a quasi-cleavage fracture mechanism, whereas that of Sn3Ag2.8Cu-0.1Ce is dark and rough, and has a fibrous pattern, and is a ductile fracture mechanism; the fractography of Sn3Ag2.8Cu-0.1Ce includes more compact and more uniform dimples than that of Sn3Ag2.8Cu, this is cause of the trace amounts of Ce refining the microstructure; brazing with the Cu substrate, the diffusion layer of Sn3Ag2.8Cu solder with Cu substrate includes more irregular IMC compared to Sn3Ag2.8Cu-0.1Ce and Sn37Pb..

Creep and High-Temperature Isothermal Fatigue of Pb-Free Solders

The creep resistance of Sn3.9Ag0.6Cu Pb-free solder alloy is compared to that of the baseline eutectic Sn37Pb solder at comparable homologous temperatures. Sn3.9Ag0.6Cu is significantly more creep-resistant than Sn37Pb solder. The isothermal cyclic mechanical durability of Sn3.5Ag and Sn3.9Ag0.6Cu Pb-free solder alloys are presented and compared to that of the baseline eutectic Sn37Pb solder at comparable homologous temperature. Cyclic mechanical tests are performed at high temperature at various strain-rates and load levels, using a thermo-mechanical-microstructural (TMM) test system developed by the authors. The data is analyzed using standard power-law durability models based on 50% load drop, using cyclic work and cyclic inelastic strain range. The durability curve of Sn3.9Ag0.6Cu Pb-free solder is found to have the largest slope, followed by the Sn3.5Ag solder and finally the baseline Sn37Pb eutectic solder, under the test conditions investigated. At a homologous temperature of 0.75, Sn3.9Ag0.6Cu shows the best durability, while Sn3.5Ag and Sn37Pb have very similar durability performance according to damage relations based on either work or inelastic strain range. The damage propagation rate is also estimated for all three solders, based on the load drop rate, and plotted vs. cyclic work and cyclic inelastic strain range.