A mission-critical industry can be defined as one in which product failure can be catastrophic: Threatening life or critical infrastructure, causing unacceptable collateral damage and resulting in liability for OEM and/or EMS. Generally included are the military/aerospace, aviation, medical and automotive industries. To confidently use lead-free in those high-reliability applications, especially considering the EU drive to impose RoHS on some areas currently exempt, it seems prudent to step back and determine where we are, how we got here and what remains to be done.

How We Got Here

RoHS was implemented in the consumer sector before hi-rel industries for two reasons. First, consumer goods made up most landfills. But, more importantly, it was initially assumed the life span for items like cell phones would give us approximately 10 years to encounter and solve any lead-free related problems before moving into mission-critical areas. Life span, however, was greatly overestimated. Cell phones, computers, etc. are now replaced sooner and with greater frequency than originally anticipated. With this reduced time frame, we have neither seen the full extent of lead-free reliability problems, nor developed means to fully combat those we have. Can we really proceed to mission-critical areas with full confidence?

Concerns

Lead-free concerns focus on two major areas: Component failure and solder-joint failure. The distinction we are making is that failure can occur either within a component itself or in the connection of component to board. Complex components, such as stacked 3-D or WLP-CSPs, can be stressed by having higher-temperature reflow cycles during their assembly and packaging, as well as during board assembly. 

Failure can also occur in passive components cited by J-STD-075 as thermally sensitive, but not accommodated in thermal profiling. And finally, it can occur in the usual suspect--the solder joint connecting component to board.

The risk areas overlap in that failures usually have the same cause: Overheating from higher temperatures used in lead-free reflow. Whether from too fast or too hot a ramp phase, too much time above liquidus (TAL), too great a peak temperature or too rapid a cool-down, failure can result. Precise processes and planned thermal profiling/verification is paramount in preventing such failures.

Component Failure

Lead-free component failure is insidious because it often appears in the field after successful test and inspection. That's one reason IPC issued the aforementioned J-STD-075. It shows that even passive components such as capacitors (actually, especially capacitors because of wide variation in material and form) can be overheated in lead-free reflow, pass functional test, yet fail in usage, bringing down the entire product. J-STD-075 recommends component manufacturers clearly mark such passives as thermally sensitive, so they can be accommodated during process design and subsequent production. 

We had already begun addressing lead-free issues of the complex components that are increasingly multi-functional, miniaturized, and subject to reflow during their build. Efforts included adjusted profiling guidelines, component shielding during reflow, more advanced test and inspection (over, under, around and through), and if necessary, hand soldering. But, because RoHS was only enacted in 2006, we still don't know whether those efforts alone will prove adequate for long-term mission-critical applications. The military, for instance, has said it wants a minimum of 15 years before first-repair.  medical wants at least seven (usually more); and even automotive wants components to last seven to 10 years [1].

In 2009, David Cavanaugh, Director, Corporate Component Engineering at Benchmark Electronics, said it will take four to five more years before we are able to have the confidence in lead-free that we do in tin/lead [2]. If he is correct, that would mean a minimum of at least three more years to go.

Solder Joint Failure

While solder joint failure may seem obvious to spot, it actually isn't. Joints can "look good" during test and inspection, but still be brittle, making them more likely to fail. In addition, cratering, head-in-pillow, graping and a whole range of similar defects can occur with BGAs and the highly functional components. Many technical papers have discussed these problems at length. Key to prevention is consistently cited as controlling the temperature/time of the reflow process: Determining correct thermal profile, maintaining it during reflow and verifying the oven throughout high volume or changeover situations. 

Changeover, however, becomes its own issue. In low-volume production, EMS providers may move from one job to another without verifying that the profile requirements are the same. All lead-free is not equal. And in high volume arenas, repeated runs are often performed to complete an order without verifying the oven remains in spec [3]. Trends as simple as dirt buildup on fan blades can knock an oven off-spec. The temperature may be correct in one zone, but not necessarily in another, or in one part of a zone, but not at the level of boards on the belt. Verification is vital.