Fictions of Pb-free surface mounting: new components have no SnPb history, making it important to keep process chemistries up to date.

The thing I love most about surface mount is that it never gets boring. The drive for miniaturization continually requires smaller footprints and higher I/O densities. As soon as production technology refines the process for a new package to the point where yields are respectable, a smaller, more challenging product comes online. When I started building circuit boards, in 1990, our advanced manufacturing group was working on developing assembly processes for 0.020" pitch QFPs and 0603/0402 components. Fast forward 17 years and today's advanced manufacturing groups are in wafer-level chip-scale package, dual flat no-lead (DFN) and package-on-package (PoP) devices. And 01005 chips (one-sixth the footprint of those seemingly tiny 0402s) are in full production in some high-end consumer products!

The miniaturization road has been a fun one, but while the supporting technologies--packaging, PWB fabrication, assembly gear and process control tools--have evolved to support the efforts, one thing remained constant: the SnPb solder used to put it all together. Engineers intuitively understood the behavior of SnPb solders, and they could rely on that intuition to manage and improve mass-soldering processes. The SnPb alloy was the cornerstone of those processes. While the switch to Pb-free alloys changes some of the fundamentals of how we build circuits, it can easily cast doubt over all aspects of the new process.

A few months ago we examined some Pb-free wave soldering rumors. Now it's time to tackle similar rumors that have circulated regarding SMT:

Pb-free solder pastes do not print as well as SnPb. Fiction. To the contrary, newer Pb-free pastes often demonstrate better print properties than their SnPb predecessors. With no-clean pastes, material behaviors like stencil life, response to pause, transfer efficiency and volume repeatability now meet or beat those displayed by best-in-class SnPb products against which they are benchmarked. Product developers now have about five years' experience when it comes to working with SAC alloys, and have developed some impressively clever techniques to balance paste performance properties.

While it is true that many first- and second-generation Pb-free solder pastes were deficient in numerous areas, available commercial products are vastly improved. When reviewing print capability studies, note the date of publication or the generation of paste that was the subject of the study. Anything published more than two or three years ago likely reflects a technology state that has since been surpassed.

Pb-free solder pastes produce more voids than SnPb. Fiction. By volume, solder paste is approximately 50% flux and 50% metal. Voiding is predominantly a function of the flux. If certain volatile components of the flux do not completely outgas before the metal melts, voids will be created. But just like print properties, Pb-free pastes' voiding properties have come a long way. A study in CIRCUITS ASSEMBLY, January 2007, compared voiding properties of two leading SnPb and Pb-free pastes on a variety of BGAs. Voiding rates observed in the Pb-free system were no greater than in the SnPb control. In most cases, slightly fewer voids were seen in the Pb-free products.

When moving to Pb-free reflow soldering, this basic principle holds true: To burn off volatiles and minimize voiding, some solder pastes perform better under soak profiles in the reflow cycle. Others perform just as well under ramp profiles. It depends on the constituents in the flux. The process engineer needs to understand the voiding sensitivity of a solder paste to its reflow profile prior to developing production processes.

More paste will be needed in intrusive reflow applications due to Pb-free products' lower metal content. Fiction. The 50:50 ratio of flux to metal remains constant in both SnPb and Pb-free solder pastes. The confusion stems from the way metal loading is expressed in product descriptions. It is indicated as a percent weight, not percent volume, because percent weight is a more precise measurement.

SAC alloy is less dense than SnPb alloy, so a certain volume of SAC weighs less than the same volume of SnPb alloy. To maintain the 50:50 ratio of flux to metal, the same volume of SAC solder powder equates to a lower weight percent (88-89%) in the final product than in SnPb (89-90%). Be assured that printing the same volume of paste from the same stencil aperture will provide the same amount of metal to fill the barrel and create the solder joint.

While on the topic of intrusive reflow, here's another basic principle that carries over directly from SnPb: A paste with a propensity to hot slump is highly undesirable, as it can drip through the holes in the PWB, leaving less metal in the hole to form the solder joint and more in the reflow oven to clean later on. When looking at reflow profiles to minimize voiding, hot slump properties--which can also create solder bridges and mid-chip solder balls--should balance the considerations.

If Sn-rich solders attack stainless steel, are nickel stencil foils and squeegees necessary? No. I have to admit, I chuckled when I heard this one. But it is a good illustration of how such a considerable change can lead us to question some of the basics. Concerns over Sn-rich solders attacking stainless steel apply only to the solders in their molten states; thus it is a concern only for wave-soldered and reworked hardware. Pb-free solder pastes have been printed with stainless stencils and squeegees for years with no signs of accelerated erosion.

Even back in the SnPb era, I was a strong proponent of keeping process chemistries current by requalifying soldering products every two to three years. When the most modern materials available aren't used, the process engineer can waste a lot of time resolving issues that have already been solved by next-generation product. Although the focus of this column is on solder paste, the same principle applies to flux, adhesives, cleaners, rework products, even conformal coatings. Understandably, it is difficult for high-reliability assemblers to convert process chemistries because of the high costs of qualification. However, for the majority of assemblers, costs incurred by the qualification of new materials are quickly recouped by the mitigation of defects that result from the change.

This months' lesson learned? It is more important than ever to keep process chemistries up to date. Nearly all Pb-free soldering products have improved by leaps and bounds during the past three years, and the vast majority of new products in development are Pb-free. Those 01005 chip components mentioned earlier were never soldered with SnPb. They entered production with a Pb-free solder paste that was designed to repeatably print those 0.007" feature sizes and reflow them in air. And the products being designed for WLCSP, DFN and PoP? All Pb-free. It's conservative to predict that by the end of 2008, Pb-free pastes will completely outperform SnPb in every performance category measurable. Except for maybe one: spread on copper. That's one of the unfortunate facts of Pb-free SMT that we'll explore next month.

Chrys Shea is an R & D applications engineering manager at Cookson Electronics (cooksonelectronics.com); [e-mail]. Her column appears monthly.

[ILLUSTRATION OMITTED]