By Tom Adams
Lead-free processing is represented in the IPC/JEDEC standard that sets moisture sensitivity levels for plastic-packaged components. The revised standard, J-STD-020D, is used by component manufacturers to expose a given component type to a specific temperature/humidity environment and then test the component. A component manufacturer that adheres closely to J-STD-020D will be able to guarantee that components will survive reflow at a given temperature without significant internal defects.
Nothing in the revised standard applies directly to military or aerospace electronics systems, but the influence of the standard will be felt because of the large number of commercial off-the-shelf (COTS) components that will be used. An engineer designing a military or aerospace electronics system can have greater confidence that components coming in the door, whether they are leaded or lead-free, will not be damaged during reflow. The influence will be global, since essentially identical versions of the revised standard have been or are being adopted in Europe and Asia.
Steve Martell, a manager at Sonoscan Inc. in Elk Grove Village, Ill., and chairman of the IPC group revising the standard, explains why the revision was necessary. “The molding compounds that are used for components that will go through the higher-temperature, lead-free reflow process typically absorb moisture more slowly than components designed for lower-temperature, leaded solder reflow. Both types of components can be damaged internally when moisture flashes into steam during reflow, but the higher temperatures seen by lead-free molding compounds make them significantly more susceptible to damage. Therefore, the lead-free molding compounds must be more resistant to such damage or the components have to be classified at a moisture sensitivity level with less floor life time.”
Establishing the moisture-sensitivity level for each of hundreds of types of components required a great deal of empirical work, Martell observes. Carried out at NXP, Agere, IBM, Intel, and other major component makers, this work consisted of baking incoming components, soaking them according to J-STD-020D specifications, placing them loose on a bare board, and running them through three reflow cycles. Three trips through reflow were needed because in production some boards are reflowed twice—double-sided boards, for instance, and boards that are reflowed once for leaded components and once for lead-free components. Boards may additionally undergo rework. After reflow, the test components are imaged on an acoustic microscope, tested electrically, and often physically sectioned.
A key change in the standard is in the maximum temperature that a component can tolerate. Under the old standard, the component manufacturer might specify that a lead-free part would survive reflow at 260 degrees Celsius, plus or minus 5 degrees. That 5-degree temperature range created some ambiguity, Martell says. “Suppose the part had been tested by the manufacturer at 255 C, but was reflowed by the user at 260 C? The part could fail even if it met the requirements of the standard.
“The standard gives one ‘classification temperature’ for each moisture sensitivity level. This temperature is dependent on the thickness and overall volume of the component. If the classification temperature is 260 C, the component manufacturer will test it at that temperature or even, to avoid problems, a few degrees higher.”
Section 6 of the revised standard spells out the distinctions between unacceptable defects and marginal defects. A small delamination in a relatively harmless location, for example, might be acceptable. A small delamination might never expand in a direction that would intersect a wire or a wire bond. But a delamination that runs the entire length of a lead finger, and thus forms a convenient pathway for outside moisture and contaminants to find their way to the bond pads on the chip, is unacceptable.
Nothing in the revised standard applies directly to tin whisker growth in unleaded solders, Martell observes. He notes that JEDEC already has guidelines for observing and managing tin whisker formation, and that other standards organizations around the world are writing guidelines as well.
Industry consultant Ron Lasky of Dartmouth College in Hanover, N.H., anticipates that the J-STD-020D standard will be beneficial to both manufacturers and users of components. “The concern is that the higher reflow temperatures of assembling lead-free solder can cause moisture sensitivity fails in components that wouldn’t happen in leaded assembly because you’re at a temperature that’s about 20 degrees cooler,” he says. “What they’ve done is change the standard to be sensitive to that. The good news is that the standard will—if people obey it—likely cause much less failures due to the sensitive device issues at the higher reflow temperatures.”
Lasky points out that some lead-free consumer products have been in use for more than five years, and the overall reliability of these products appears to be fairly good so far. “There have been really few documented fails of lead-free products beyond the noise level that you’d expect in any product,” he notes.
Sonoscan’s Martell explains that the influence of the standard on military and aerospace designers and assemblers may be somewhat complex. Because a given board is likely to have components whose volume ranges from large to small, and because these variously sized components are arranged at various distances from the edge of the board, it is somewhat of a challenge to avoid over- or under-heating a given component. A common problem is applying adequate heat to a relatively massive component near the center of the board without roasting a much smaller component near the edge of the board.
Using the specs in the standard to do a better job of avoiding problems, a designer might decide to use only lead-free components on a particular board, Martell adds. Actually, this decision might become inevitable as suppliers discontinue leaded components. Alternately, the board might have both lead-free and leaded components. “You might want to reflow the lead-free first at the higher temperatures,” he says, “and then add the lower-temperature passive components, or whatever components that you can still get in a leaded variety.“
Lasky believes that the revised moisture-sensitivity standard gives users a level of assurance that they will avoid problems in moisture-sensitive devices. He also thinks that there are limits to this assurance. Military systems may sit in storage for years, and then be put into use when needed. Expecting lead-free systems to remain in a warehouse for a decade or so, and then be put into use in a harsh environment may not be realistic, because the overall experience with lead-free components, although fairly good, is limited.
As more experience is gained with lead-free systems, Lasky expects there will be lead-free failures that will be well publicized. “I don’t believe it’ll be a the-sky-is-falling type of thing,” he says. “It’ll be isolated. With leaded solders we’re still learning things that we didn’t know about.”
Lasky offers an admittedly subjective scenario to put the likelihood of lead-free component failure into some sort of context. “Lets say there are a billion laptop computers in the world, and over the next three years 100 million of them will fail for one reason or another. Of the 100 million fails, 10 million will be due to hard drives, 20 million will be due to plugs that don’t work, 30 million will be due to keyboard failures. I think you’d find 114 fails that have something to do with lead-free solder. I’m going to lay awake at night because I’m going to get some fails due to lead-free solder? I’ve got all these other fails that nobody’s even talking about, because they don’t have anything to do with solder.”
More seriously, Lasky observes that, if he were a senior engineer in charge of military reliability for mission-critical systems, he would still be “very cautious and concerned about the possibility of fails in mission-critical products that I got as a COTS item.” But such an engineer can do many things to greatly improve long-term reliability—working with suppliers, for example, to use nickel flash on the copper to minimize Kirkendahl voids. “We know enough to know what kinds of things we can do that will most likely give you the long-term reliability,” he says.