News
Pb-Free Solder Joint Reliability
In God we trust; all others must supply data.
"You give up something for everything you gain ... So pay for your ticket and don't complain." – Bob Dylan
The RoHS Directive and Pb-free may be a wrong direction for a number of reasons, some valid, others questionable. The same can be said about why RoHS may be beneficial. At the heart of the debate is the environmental question upon which RoHS was promulgated. The RoHS Resistance Movement, as I call it, is quick to note the oft-cited University of Stuttgart study1, which asserts that the harm RoHS causes to the environment outweighs the benefits. While there are some gaps in that study and a follow-up would be great, it is nonetheless data.
But another RoHS Resistance position is that Pbfree solders, component and surface finishes will produce less reliable solder joints than SnPb joints. We've heard numerous times a variation of, "I just know that Pb-free will be less reliable ...." This begs the question: How do you know this? Do you have supporting data, or did you pull this assumption out of the air?
The Resistance has pointed out that "expert opinion says that reliability will be an issue and that the replacements for leaded solder are simply not as reliable as their predecessors..."2 Who exactly are these experts, and where did they conduct and publish their research? As we say at ITM: In God we trust; all others bring data.
Let me make clear that I do not agree with RoHS and how it has been foisted on the populace. I wish the RoHS Resistance well. I agree with Thomas Jefferson, who said, "Dissent is the highest form of patriotism." But we have to be scientific. (To be fair, some proponents of Pb-free purport its superior reliability compared to SnPb. Again, let's see some real, noncommercial data.)
In 30-plus years in this industry, I've seen quite a few field failures of SnPb interconnections, but that doesn't mean that SnPb should be branded unreliable. I saw Pb-free in use long before RoHS. I remember a Unisys application in the late '80s for which CCGAs and other components were assembled to a motherboard with an SnBi alloy. SnAg alloys have for years performed reliably in a number of high-temperature applications (e.g., under-the-hood automotive) predating RoHS. Then there's myriad Japanese Pb-free consumer electronics that have been manufactured for well over five years. The Japanese are not exactly known for risking their reputation on unreliable materials and design (although Sony's self-destructing laptop batteries bear watching).
As a consultant and columnist, I must be objective and pragmatic. One engineer who has researched and published is Dr. Jean-Paul Clech of Electronics Packaging Solutions International (jpclech.com). Dr. Clech takes an honest, objective, scientific and, yes, pragmatic approach to the matter. He begins by stating what should be obvious: "The important question is not whether Pb-free is better or worse than SnPb under accelerated conditions, but whether product reliability requirements are met for SAC assemblies."3 Design for reliability:What a concept!
Through his own tests and researching countless other studies, Dr. Clech has studied SAC alloys extensively. Dr. Clech asserts that Pb-free solder joint acceleration factor modeling, when combined with actual test failure data, provides a fairly accurate and expedient way of assessing solder joint reliability. From these values, test results can be extrapolated relative to field conditions and provide an estimate of expected PCBA reliability. The alternative, of course, is to "wait and see and take a chance," but we saw problems with that approach in SnPb assemblies.
Based on test data and models, Dr. Clech has brought forth some intriguing information. The industry trend is toward SAC compositions with lower silver content. Dr. Clech found that lower silver content leads to faster creep rates and lower thermal cycling life. However, the improved ductility may help under mechanical loading. He observed a strong correlation between SAC acceleration factors obtained from compact strain energy or FEA modeling and actual test results. (Take care when using algebraic models, he cautions.) Modeling is important when setting reliability expectations, but must take into account the operating conditions of the application.4
The operative word here is application. Once again, as is almost always the case in electronics assembly, things tend to be application-specific. From a comparative perspective, Dr. Clech found that the SAC-to-SnPb life ratio decreases with 1) increases in cyclic strains (or temperature swing), 2) dwell times and 3) mean temperature. The bottom line: Blanket statements about Pb-free solder joint reliability need to be put in context. The most important factor is whether product reliability requirements can be met and thus, whether the Pbfree interconnection/product life is sufficient for the application and its environment. SnPb field failures, while relatively few, are well understood, usually in relation to poor workmanship and a lack of DfR. As such, many SnPb assemblies may have been over-designed, leaving room for more optimized Pb-free designs.
As was the case once upon a time with SnPb solder, as more data become available for Pb-free solder alloys, we will be able to develop comprehensive lifetime correlations. Some applications have already benefited from conversion to Pb-free because of the product environment. At this juncture, the life of mixed assemblies (Pb-free and SnPb) is difficult to predict because of the variable non-homogenous microstructure of the resulting solder joints. Tin whiskers, Kirkendall voiding, tin pest and other issues need to be further understood and put into proper perspective (don't panic). But we will continue to learn as time goes by. In the meantime, remember, we're all in this together.
On behalf of the industry, special thanks to Dr. Jean-Paul Clech of EPSI International for his excellent work.
References
N. Warburg, IKP University of Stuttgart, Department Life Cycle Engineering, warburg@ikp2.uni-stuttgart.de.
J. Burke, rohsusa.com, April 2006.
J.P Clech, "SnAgCu Assembly Reliability Trends and Acceleration Factors," SMTAI Proceedings, September 2006.
R. Zhang and J.P. Clech, "Applicability of Various Pb-Free Solder Joint Acceleration Factor Models," SMTAI 2006 Proceedings.
Phil Zarrow is president and SMT process consultant with ITM Consulting (itmconsulting.org); itm@itmconsulting.org. He still bears the scars, physical and mental, of reflowing convection/IR ovens. His column appears bimonthly.
