Lead-free mandate plumbs new design challenges
Initiatives for getting the lead out of electronic products will affect design, debugging, components, testing, and even reliability.
By Bill Schweber, Executive Editor -- EDN, April 18, 2002
|
Most design engineers don't worry about production issues, such as soldering, testing, reworking, and basic pc-board reliability—and with good reason. The production and manufacturing engineers, OEMs, contract manufacturers, and related specialists have worked long and hard to make sure that, for most situations, board-component attachment, soldering, drilling, and plating-through are not a problem. They have achieved this goal as component sizes have shrunk, line widths and spacing have dwindled, ICs have placed pins under the packages, lead pitch has tightened, and lead count has increased. In short, they have met the challenge and given circuit and system designs a break from one big area of potential headache.
But this bliss may be ending. Worldwide initiatives to reduce and even eliminate the use of lead and its various compounds in electronics have developed traction in Japan and Europe, and the United States is seeing evidence of this trend as well. The lead-free movement brings new challenges and unknowns to design, and you may have to be ready to meet them.
Lead-free in a state of flux
Lead's usefulness parallels its potential danger (see sidebar "A brief history of lead"). In 1994, a European consortium known as IDEALS (roughly corresponding to the acronym for Improved Design Life and Environmentally Aware Manufacturing of Electronic Assemblies by Lead-Free Soldering) focused on getting lead out of electronic soldering and began investigation lead-free alternatives (see sidebar "How much lead to get out?"). By 1998, the European Commission introduced a draft called the WEEE (Waste Electrical and Electronic Equipment) directive asking, among other things, for a ban on lead in electronics by January 2004. This initial target date has been pushed back several times. It is currently set for 2006, but even this date may be delayed.
Leading active-component suppliers have promoted initiatives to eliminate lead from their semiconductor products, exemplified by the July 12, 2001, joint announcement by Infineon Technologies AG, Philips Semiconductors, and STMicroelectronics. Of course, like most official directives, this one contains exceptions and loopholes. Military, medical, storage, and some other types of equipment are exempt for now, for example.
In Japan, the lead-free effort has had early momentum and tangible results. For example, in 1998, Matsushita (Panasonic) introduced the MJ30 lead-free minidisk player and was soon shipping 40,000 units per month. The same year, JEITA (Japanese Electronic Industry Technology Association, www.jeita.or.jp) developed a road map for lead-free soldering. Many Japanese companies have publicly announced their intentions to go lead-free in specific consumer-product lines in the next few years, seeing the move as an environmental and marketing plus.
Why is it your problem?
Even with today's leading-edge designs, you normally don't have to worry too much about manufacturability and reliability if you do your homework or use a properly qualified reference design. Your hardware should work consistently if you take the time to do proper simulation of the layout, tolerance analysis of component parameters, and assessment of timing margins and if you employ proper power distribution and bypassing, watch localized and overall thermal analysis, and follow good design practice. Yes, there may be problems due to noise, I/O issues, misassumptions about components, and other hard-to-foresee nastiness, but fundamental reliability is usually a starting point in today's design. (Software reliability is another story, but it's not relevant to the lead-free situation.)
But once you decide to go lead-free, you face design-related issues in component availability, design rules, production and test problems, long-term reliability, and factory and field rework and repair. For a few of these challenges, the answers are new but known; for most, the questions themselves are unclear, and the answers are even harder to see.
Getting the parts list for your bill of materials is the first challenge. A design's bill of materials is not just a list of ICs with little else. Currently, for a conventional design, you can choose from a wide array of vendors for active components, passive components (a typical design contains 30 to 50 passives per IC), switches, optical and display components, and more. Although vendors are offering some lead-free components (Figure 1 and Table 1), you'll certainly be more restricted than you have been. You may find only one source for a part, a part that is not quite the one you want, or perhaps no source at all for that special component you need. If your bill of materials includes an IC that is not an off-the-shelf catalog item, as most designs now do, you have to work with the vendor to ensure that you can get lead-free packaging for it. Your customary leadtimes may also change.
Your established pc-board-design rules also need to change. Currently, industry guidelines govern component lead-pad and land size, track width and spacing, via and through-hole dimensions, and similar factors to ensure manufacturability and reliability. But, the physical characteristics of any solder include subtle factors, such as its ductility and elasticity. In addition, the local heating of component leads and their pads causes some thermal expansion during operation, which tin-lead solder accommodates and matches. The lead-free solder substitutes now being deployed have different thermal and physical parameters and so may develop minute cracks between the component lead and the pc-board pad due to thermal cycling. It's even possible for thermal changes to cause the lead and solder to pull the cladding-pad material from the pc board, so you may need larger lands to give you increased grip-surface area between the pc-board cladding and the pc-board substrate itself. The thermal issues worsen when a component uses the pc board's copper as its heat sink, with a thermal path from the die to the board through the component leads, as is common in today's designs, or a thermal pad under the IC to provide a larger path to the copper underneath.
Your test strategy may need to adapt as well. Bed-of-nails test fixtures apply carefully controlled forces to the test probe, so it makes proper contact with the board-under-test solder bumps; this force is based on the yield, hardness, and resilience of the bump. The increased hardness and brittleness of nonlead solder bumps may instigate joint failure, so you'll need to consider revising your fixtures or using more built-in-test and JTAG-like techniques to reduce the need for such fixtures.
For advanced designs based on hybrid-fabrication technology, the situation is even more complex. These designs often build up system circuitry using a sequence of assembly and solder operations, starting with a higher temperature lead-based solder and moving down to lower temperature solder-connection steps, in one or more extra stages. This design-and-assembly plan depends on the well-known melting points of lead-based solder alloys, down to the common eutectic combination, which melts at 183°C. With lead-free solder, those well-known markers become unavailable, and the flexibility you once had may be gone.
If you think your job is done when you turn your lead-free design over to production, think again. The well-established reflux- and wave-soldering processes have many success factors, and two are especially noteworthy. One is the temperature and heat energy that the board must see to melt the solder, and the other is the profile of the heat-up and cool-down phases that occur before and after the solder reaches the melting point. Most equipment in use can provide the necessary heat, even though you have to raise the setpoint temperature 20 to 40°C for the lead-free soldering. However, whether this equipment can also properly control the overall ramp-up, soak, and cooling cycle profile is something you and your manufacturing engineers need to investigate. Changes in flux chemistry and board-washing techniques—factors that normally do not concern design engineers—may reflect back into component selection, how components are attached, or other former nonissues.
When your company (or contract assembly provider) does any postsoldering manual rework—whether to touch up boards, install special components, or implement last-minute design fixes—all the rework must use the same lead-free solder. And if you do any field repair at company-service centers (yes, many electronic devices still get repaired, not just replaced), the lead-free mandate must be in place there, as well, with right lead-free alloy, suitable desoldering and soldering stations, and personnel training. Note that, in general, you cannot mix different lead-free solder formulations on the same joint.
But your concerns don't stop there. Although our industry has uncountable hours of field and factory experience, thousands of tests and long-term studies, plus millions of reliability-related data points on the use of lead-free solder and manufacturing issues, little similar data and knowledge exists for lead-free designs. In addition, virtually all of today's leaded solder uses one specific alloy composition, so data and knowledge from different sources are generally comparable. The lead-free domain, however, has different alloys and thus different potential problems. Therefore, the data collected for one specific alloy may not be useful for understanding the immediate production issues and longer term reliability of other alloys.
Component reliability is another issue. The higher melting temperatures of the lead-free solders that are coming into use mandate components that can withstand these increased temperature stresses and soak time of the soldering process. Life-test data for many components at these higher temperatures is less comprehensive than it is for processes using lead; many components simply aren't available or specified for the new region.
Long-term-reliability factors are the greatest unknown for the designer. The other issues cited above—new design rules, component-selection constraints, and factory manufacturability and rework—can be handled with careful planning and hard work, but at least these problems are known. Will the usual suspects—software bugs and marginal hardware design—cause subtle or intermittent problems and system crashes or glitches? Or will new failure modes, such as solder-joint crack growth, stress induced in components during the higher temperature soldering process, mysterious solder and board ailments that we don't yet know about, or something else be the culprit?
Some historical perspective may give you cause for concern: The current high level of hardware reliability and manufacturability came about not by a smooth, linear process, but as the result of research, experiments, and factory and field experience. Uncovering the secrets of optimum component structure and soldering included the often-painful investigation of lead, solder-alloy, and pc-board material science; process-temperature profiles; joint wetting; flux chemistries; and cleaning solutions. Solder- and materials-related problems, such as stress fractures, corrosion, fungus-like growths, and more were uncovered and solved through many late nights of investigation, factory and field data, and anecdotes—at considerable cost.
What should you do now? There is no easy answer; the industry has yet to standardize even the most likely alternatives to tin-lead solder. Whether you have a mandate to design a lead-free product or are considering doing one, start looking now at all your critical components, design rules, fabrication processes, component engineering, and reliability records. Try to avoid redoing R&D that others have done already, and be prepared to work with manufacturing-process specialists and consultants who have experience and the scars to show for it. Most important, keep track of industry developments via professional or industry associations, as well as press reports. Table 2 shows some useful resources, references, and sources of information.
| Acknowledgments | ||
| Thanks to Karl Seelig, vice president of technology at AIM Inc; Barry Marsh, vice president of marketing at Actel Corp; and Luc Petit, a specialist at STMicroelectronics, for their insight. | ||
Acknowledgments
Thanks to Karl Seelig, vice president of technology at AIM Inc; Barry Marsh, vice president of marketing at Actel Corp; and Luc Petit, a specialist at STMicroelectronics, for their insight.
| Author Information |
 Executive Editor Bill Schweber was often told to "get the lead out" in his earlier years. You can reach him at 1-617-558-4484, fax 1-617-558-4470, e-mail bill.schweber@cahners.com. |
