GENERAL CONCLUSIONS / RECOMMENDATIONS

A review of material properties of Sn3.5Ag and SAC solder alloys was conducted. The review is not exhaustive or complete, and is limited to publications in the English language. A lot of scatter was observed in the data although some trends are clearly visible, as discussed in the "Conclusions" sections of the SnAg and SAC text above. More general conclusions and recommendations are:

• Discrepancies were noticed between shear and tensile data.

• The applicability of the Von-Mises yield criterion has not been demonstrated for the alloys of interest. For example, the Von-Mises criterion does not seem to apply to Sn3.5Ag in the low stress regime.

• Discrepancies were also noticed between bulk solder data and data obtained from solder joints of electronic assemblies.
• Since mechanical properties are strongly influenced by the micro-structure, it is important that the latter be described in a quantitative manner for proper interpretation and use of the data.
• The microstructure itself, for example the size of the b-Sn globules in Sn-Ag and SAC alloys, depends very much on the specimen cooling rates. While some studies report on both, others do not mention either one, which makes it difficult, if not impossible, to put the data in perspective with the results from other experiments.

• For reference purposes, standards (e.g. JEDEC standards) for the assembly of electronic circuit boards recommend maximum cooling rates of 6°C/sec. This is relatively slow cooling when compared to rates of 50-150°C/sec when water-quenching cast alloy laboratory specimens.
• For the purpose of developing accurate constitutive models for solder joints of electronic assemblies, the initial micro-structure of test specimens should have features (e.g. b-Sn dendrite globules) of similar size as identical features measured in real solder joints after assembly. This is not only a function of cooling rates but also a function of test specimen and solder joint sizes (for example, in terms of volume, a flip-chip solder joint is about 200 times smaller than a conventionnal BGA solder joint).

• Very little data is available at stress levels below 10 MPa, the stress range of interest under many service conditions. While creep tests conducted at low stress can be time-consuming, this process can be accelerated by testing at higher temperatures.
• The results of most mechanical tests emphasize secondary or steady-state creep deformations or strength. Other deformation modes, such as initial deformations, rapid plastic flow and primary creep, are rarely reported on because they appear less significant under high stress conditions. However, these deformation modes need to be investigated further since they may become more significant under rapid thermal cycling conditions with high ramp rates and/or short dwell times.
• Data from a given publication are rarely bounced against data from other sources. Their applicability to other conditions or to solder joints of electronic assemblies requires further analysis and investigations.