Copper dissolution, initiated by Pb-free alloys, may threaten long-term reliability.

In all the writings about issues caused by RoHS legislation and the subsequent need to find Pb-free solder alternatives, one issue seems to have lacked publicity but is potentially disastrous for safety-critical products: copper dissolution.

During the soldering process, CuSn intermetallic compounds form at the boundaries between component terminations and the solder joint, and also at the pad-to-joint termination. In Pb-rich solder, not much copper is absorbed into the intermetallic; the pad and trace retain much of their post-bare board fabrication shape and size. The trace and pad are therefore in a strong position to maintain a reliable connection between the various parts of a circuit.

Many Pb-free solder alternatives are predominantly SAC alloys: tin, silver and copper. For cost reasons alone, it makes sense to limit the amount of silver in any given alloy. That means these alloys are rich in tin and copper, which in turn facilitates the ability to form CuSn compounds in large volume. With a large amount of tin and proportionally less copper present in the solder alloy, the tin needs to acquire more copper from somewhere to create a stable CuSn compound. This copper can come only from the board traces and pads.

When temperature and time at high temperature are raised to correctly solder the Pb-free alloy, the joint is more likely to draw more copper from the surrounding pads and traces. So, by using Pb-free alloys, we are deliberately increasing the risk of ruining our boards.

Wave soldering requires higher temperatures than does reflow; the occasional rework or hand soldering can also introduce very high temperatures. Selective soldering acts like a miniature wave soldering process and carries high temperature penalties. These technologies, therefore, are coming under the greatest scrutiny. This does not mean reflow soldering can be ignored as a source of copper dissolution, but we need to concentrate on wave varieties first.

High reliability products are, for now, covered by exemption rules and may use SnPb solders. But for how long? As long as they use lead, I will continue to fly airplanes, or until we find ways of avoiding copper dissolution issues and other scary problems.

A government-backed project underway in the U.K. at the National Physical Laboratory (npl.co.uk) involves all aspects of manufacturing from PCB manufacture through component design and selection to product assembly. Many of the world’s leading equipment and materials suppliers are contributors. (Bob Willis, one of the project leaders, has conducted much research, and the photographs shown are courtesy of him and the NPL.) The project goals are to understand copper dissolution mechanisms and develop or modify processes, and perhaps equipment, to minimize or eliminate the problem.

What does it look like? Some of the photos reveal the excesses of copper dissolution, demonstrating sometimes so little copper remains on the board that interconnections are tenuous at best. We do not yet know what the long-term reliability of such joints may be.

Figure 1 shows part of a selectively soldered 96-pin connector. The lack of copper in the lands and into the hole barrels is pronounced. Figure 2 shows a microsection of a joint previously assumed to be as good as any wave soldered joint, but note the lack of copper at the knee where the pad meets the hole barrel. Figure 3 shows dramatic copper dissolution. The only copper showing is the remaining metal that had keyed deeply into the PCB laminate.


The NPL expects to run trials on a variety of surface finishes and alloys and hopes to obtain process parameters. How will OSP compare with Pb-free HASL? The copper on an OSP finish should be easily attacked, but the Pb-free HASL will have seen one reflow excursion before product assembly. What effect will that excursion have on the joint? Will ENIG turn out to be the best finish despite its cost (and other issues)? Will silver finishes find favor? One can make many theoretical assumptions and many companies have made some comparisons, but the purpose of this study is to define best practice guides for suppliers and manufacturers.

There are also some practical equipment-related issues to resolve such as the control of dwell times in selective soldering and control and guidance on hand-soldering stations. One suggested resolution demands use of laminates with more copper. Yet the trend in high-density interconnects is to use less copper on a laminate, not more. There would also be a cost penalty in using more copper (although no doubt copper suppliers would welcome such a move.)

The next time someone writes on copper dissolution, I hope we all have a better understanding of how to limit its effects.

Ed.: More information on copper dissolution is available from the author; Bob Willis (bob@bobwillis.co.uk); or Dr. Chris Hunt of the NPL (chris.hunt@npl.co.uk).

Peter Grundy is director of P G Engineering (Sussex) Ltd. and ITM Consulting (itmconsulting.org); peter.grundy2@btinternet.com. His column appears bimonthly.