Lead spread: Pb-free pastes do not spread like SnPb on OSP.
Pb-free solder pastes do not spread as well as SnPb pastes on OSP. Absolute fact. Figure 1 tells the story. To capture differences in spread between the alloys, we offset the stencil from the pads by about five thousandths of an inch. This degree of misalignment is not uncommon in SMT assembly. Notice that the SnPb37 paste pulls back on the pad where it was overprinted and spreads to the areas of the pad that weren't printed. Although SAC 305 paste pulls back from the overprint, it does not spread on the unprinted areas.
Like many other process engineers, I assumed the issue of lower spread was the fault of the OSP finish, but I recently learned it is just a matter of physics between the alloy and copper substrate. Lead spreads on copper. Remove the lead; remove the spread. I have not run the experiment myself, but am told that spread tests of Pb-free alloy on clean, fresh copper, and on the same surface treated with OSP, would produce similar results. It's not that the OSP hinders the spread; it's that the Pb-free alloy has no natural tendency to spread on the copper.
That's a nifty fact to learn, but OSP is one of the most popular PWB surface finishes, and we still need to assemble boards that don't show exposed copper. I've never seen a reliability study that cites "copper corners" as a root cause of catastrophic failure, but they do present cosmetic issues that often need to be addressed. As a matter of practicality, the paste deposition process must be tightened if we are to avoid copper corners when transitioning OSP assemblies to Pb-free. Registration between the stencil apertures and PWB pads is critical, and the pads must be almost completely covered with solder paste.
The registration part is easy. Stencil scaling technology has been around for years. With a stencil scaled to compensate for the dimensional variation of the PWB and a calibrated printer, the alignment between the stencil apertures and the pads should be within 0.001" or 0.002". But opening some of the stencil apertures can reasonably cause concern. When using SnPb solder paste, aperture reductions are typically employed to curb defect rates. On QFPs, reducing the width of the stencil apertures can improve gasketing to provide better print definition and limit the solder bridges caused by slump. On chip components, reducing the aperture size lessens the volumes of paste deposits to decrease the incidence of "squeeze balls," or mid-chip solder balls (MCSBs). Based on our history with SnPb, opening the apertures could result in more defects and send first-pass yields in the wrong direction.
[FIGURE 1 OMITTED]
Here comes the good news. It seems Pb-free solder pastes are more tolerant of 1:1 printing than SnPb. Despite the many downsides of Pb-free soldering, there seems to be at least this one upside:
Pb-free pastes create less solder bridges and MCSBs in the reflow process. Fact. Yes, the dull, grey cloud may in fact have a shiny silver lining, or at least a partial one. I've heard numerous anecdotes from assemblers that transitioned products from SnPb to Pb-free and reported less solder bridges and MCSB defects. We ran a benchmarking test in the lab; it showed that 1:1 apertures created more than three times as many MCSB defects on components sized 0402 to 1206 with SnPb paste than with Pb-free paste. The data were fairly clear. (It's also worth noting that the test showed more QFP bridges with SnPb, but the defect rates were so low that the results weren't statistically significant.)
This is an interesting phenomenon. We know the effect, but not the cause. I'd like to know the cause because other benefits may be associated with this behavior that we have not realized yet. Part of our job as process engineers is to understand the properties of materials and exploit them to our advantage. Perhaps the same property that limits solder bridges and MCSBs will also permit wider windows on area array and intrusive reflow overprints.
Since hot slump is a common thread in the formation of both solder bridges and MCSBs, my first thought was that the more heat-resistant properties of Pb-free solder pastes may be providing some benefit. But I'm not an expert in solder paste, so I asked Dr. Michael Liberatore, an R & D director at Cookson Electronics, for his take on the phenomenon. He says different fluxes exhibit different slump properties, so if hot slump is a factor, other factors should probably be considered also. Mike speculates that the formation of fine pitch bridges and MCSBs may be influenced by properties of the alloys in their liquidus phase. He explains that the oxide film that develops on the molten solder surface is more tenacious in Sn-rich alloys than in SnPb ones. The tougher film holds the small beads of molten solder together better, thus preventing the "breakaways" that can create MCSBs, and the flow and fusion from two adjacent QFP pads that create a solder bridge.
So what is the underlying mechanism at work here: the flux, the alloy or a combination? I'm going to set up an experiment to find out. Because we ran the benchmark test that compares SnPb and Pb-free pastes a couple years ago, I'll rerun it to maintain experimental consistency between the controls and test sample. The test sample will be hybrid paste of the same Pb-free flux combined with SnPb solder particles. If the better characteristics are a result of hot slump properties, then the hybrid paste should behave more like Pb-free with respect to the formation of MCSBs. If it's due to the oxide film on the molten alloy, then the hybrid paste should behave more like the SnPb. I'll report the results when they become available.
This month's Pb-free lesson is not one we have actually learned yet; rather, it is one we are still in the process of learning. In trying to overcome the poorer spread properties of Pb-free solders on copper surfaces, we are finding that we may be able to remove a compensation factor that became a standard practice with SnPb many years ago.
Although we have some strong indications that the Pb-free solder pastes are more tolerant of 1:1 printing, I would not advise any large-scale process changes without more data, or at least a rudimentary understanding of the factors driving these phenomena. The behavior of the hybrid paste should help us get to the bottom of it. This emerging class of hybrid solder pastes -Pb-free flux blended with SnPb solder powder--are noteworthy because their popularity is rising during this transition period. In many cases, they can provide a solution to the issue of Pb-free BGAs in a SnPb process. They can also present risks that must be thoroughly examined. More on this topic next month.
The author would like to thank Dr. Liberatore for his insights on the reflow soldering process.
Chrys Shea is an R & D applications engineering manager at Cookson Electronics (cooksonelectronics.com); email@example.com. Her column appears monthly.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Pb-Free Lessons Learned|
|Comment:||Lead spread: Pb-free pastes do not spread like SnPb on OSP.(Pb-Free Lessons Learned)|
|Date:||Jun 1, 2007|
|Previous Article:||A chance for sanity: the EU opens the door--finally--to debating RoHS and WEEE.|
|Next Article:||SAC whiskers? Metal whiskers are not a new problem. But finding them in new places can be.|