Are such BGAs backwards compatible with SAC 305?

The term backwards compatible commonly refers to a Pb-free component and SnPb paste. Efforts to resolve this issue for BGAs¹ are ongoing, and data collected so far are conflicting. However, before that issue can be put to rest, another compatibility challenge is arising.

We discussed this change in BGA ball metallurgy in a previous article.² Considering that low-Ag alloys have a melting range up to 228°C, and Pb-free assembly processes have been developed within the 217-221°C melting range of SAC 305 solders, it’s appropriate to ask: Are BGAs with low-Ag SAC spheres backwards compatible with SAC 305?

Here, we discuss examples and associated challenges of assembling BGAs with low-Ag alloy spheres with SAC 305 solder pastes.

A supplier shipped a BGA with low-Ag alloy spheres (composition “A”) without a part number change. The original SAC part was assembled successfully on a small form factor card with a peak temperature of 235°C some time prior to the change. During first-article inspection, process engineers noticed something was not the same as previous builds. The line was stopped to prevent the issue from escalating. Optical inspection and cross-section analysis revealed the joint was not forming properly (Figure 1).


After hours of analysis, and confirmation, which revealed no other process changes, the issue was brought to the supplier’s attention. It was discovered that the sphere alloy had been switched from SAC 305 to a low-Ag composition. The team returned to the assembly line for experimentation, and modified the profiles, which resulted in acceptable solder joints. The team also verified the profile change did not affect other components on the board. In the following weeks the customer also performed reliability analyses, including drop tests, and approved the changed process, starting production once again. (This is one reason why a majority of top OEMs and EMS companies are calling for unique part numbers when a BGA’s alloy metallurgy is switched. The change is necessary to understand the process impacts and better manage the process.³)

Some BGA suppliers have sent PCNs of impending changes, permitting the customer and its RoHS component engineering experts to alert production. The part in question, with a low-Ag alloy of composition “B,” was a small 40-plus I/O CSP, while another, larger SAC 305 480-plus I/O BGA was on the same board. The profile developed to ensure correct temperatures were reached for the center joints on the BGA to form proper joints also permitted the CSP’s solder joints with the low-Ag spheres to reach a peak temperature of 240°C. The package temperatures for this and all other components were within specifications. The teams then performed dye and pry tests and cross-sectional analysis (Figure 2) and sent reports and boards to the customer for approval.


As it happens, that same supplier was readying to ship to another manufacturing site. Through internal bulletins and email communications to site RoHS experts, sites were alerted to the coming change; site RoHS experts readied to engage in discussions with suppliers and customers. This helped the manufacturing site manage the change with the supplier. The customer requirement was to use the composition “A” BGA, because no alternative was available on the customer’s AVL, but while achieving a higher percentage of silver in the final joint. With some changes in stencil design and print parameters, the team was able to assemble the board and currently awaits customer direction on production release. This potential for delayed implementation due to metallurgy proliferation is echoed in several industry-supported statements.³

To return to the original question about backwards compatibility, the answer is a classic “it depends.” Industry recognizes^{3, 4} the change to low-Ag alloys may have several reliability benefits, including superior drop and shock resistance for some products, such as cellphones. As discussed, it is possible to assemble BGAs with low-Ag spheres with a SAC 305 Pb-free solder paste under certain process and assembly conditions. However, it is important for assemblers be aware of the impending change, which is best communicated through a part number change, or at least a PCN far ahead of the actual change. It is also important for assemblers realize it is a change that affects “form, fit and function,” and there are potential impacts to the assembly process and other components on the board. Proper logistics solutions such as a part number change, assembly process optimization, and reliability analysis including thermal cycling may need to be addressed to fully recognize the potential benefits of these changes.

Now the question is, Will the reflow profile solution developed over months to solve the classic backwards compatibility problem (i.e., Pb-free SAC 305 BGA in a SnPb assembly) work for a low-Ag alloy sphere BGA in a SnPb assembly without re-creating the detailed analysis? Maybe not, but we will leave that question for another day.

Acknowledgments
The author would like to acknowledge the contributions of Gaetano Covucci, Paolo Rategni, Henry Tan, K.F. Chow, Jeanette Koo, K.S. Yau and Josue Mejia from various manufacturing locations for sharing their experiences.

References

1. R. Kinyanjui, and Q. Chu, “Pb-Free BGAs in SnPb Assembly Process Project,” Apex Conference, February 2007.
2. K. Reid and G. Wable, “Changing BGA Solder Ball Metallurgy,” SMT, August 2007.
3. iNEMI, “iNEMI Members Call for Unique Part Numbers to Differentiate Ball Metallurgies on Pb-free BGA Components,” May 2007.
4. Greg Henshall et al, Manufacturability and Reliability Impacts of Alternate Pb-Free BGA Ball Alloys, iNEMI Working Group, June 2007.

Girish Wable is senior advanced technology engineer and Paul Neathway is manager, advanced manufacturing technology, at Jabil (jabil.com); girish_wable@jabil.com.