Medical Device Link.

Originally Published MEM Spring 2002

ELECTRONICS MANUFACTURING

Microelectronics manufacturing worldwide is in a state of transition: from lead-based solder to lead-free solder.

Tom Adams

Over the next several years, microelectronics manufacturing will shift from the use of familiar lead-based solders to the use of lead-free solders. Although driven by legislation in Europe, the transition in the United States will likely be propelled by market forces—namely, the desire to claim that a product or system is environmentally friendly.

This transition will create significant problems for the designers and assemblers of medical electronics systems. A year or two ago, these problems appeared perhaps even more ominous than they really are, because so little was known about the varieties of lead-free solder and their manufacturability. But real data are beginning to emerge from early users of lead-free solders and from a study conducted by the National Electronics Manufacturers Initiative (NEMI).

Reflow and Reflow Temperature

For all solders, the critical point in manufacturing is the reflow process. For tin lead, which is typically 63% Sn and 37% Pb and has a melting point of about 183°C (361°F), reflow solder-joint temperatures commonly reach 220° to 230°C (428° to 446°F). The most widely used lead-free solder is an alloy of tin, silver, and copper; the alloy used in the NEMI study was 95.5% Sn, 3.9% Ag, and 0.6% Cu. NEMI selected this alloy after a thorough study of all available data on dozens of formulations, and they are recommending that industry adopt it as the replacement for 63Sn37Pb.

It is likely that lead-free solders will be identical or very similar to the alloy studied by NEMI. The alloy's melting point is 217°C (423°F), which means that peak reflow temperatures can be expected to reach 260°C (500°F). Although most reflow profiles will peak at temperatures slightly below 260°C (this temperature was the target established during early lead-free manufacturing in Japan), the industry consensus is that planning for lower temperatures would be unwise. And the temperature that a given board experiences locally may actually be higher than this. For example, very small surface-mounted components may reach temperatures of 280°C (536°F).

Lead-free solderable finishes will be used for onboard pads and as plating on component leads, so the solder's physical properties must be suited to these tasks. The characteristics of SnAgCu are already well known: it is more rigid than SnPb and is therefore less forgiving of stresses in some situations. Also, it does not wet surfaces during reflow as well as SnPb, a difference that will result in different wetting angles. Although these differences will call for slightly altered manufacturing techniques, they do not appear to be major problems. Nor are the boards themselves likely to be a problem. The NEMI study and early production both show that FR4 boards can withstand reflow to 260°C with, at most, slight cosmetic discoloration. But less-robust board materials, such as phenolic paper, may require replacement.

The most serious problem in the transition to lead-free solder will be the components. Assemblers will need to requalify every component for the new higher reflow temperatures, largely because the higher temperatures make damage from internal moisture more likely. IPC currently recognizes eight levels of moisture sensitivity (see Table I). As a rule of thumb, an increase of 10°C in reflow temperature increases a component's moisture sensitivity by one level. Moving from lead-based solder to lead-free solder may therefore raise a component's sensitivity by three to four levels.

The major purpose of the NEMI study was to determine to what extent and in what ways components are damaged by higher reflow temperatures. The study involved more than 2000 individual components, including thin, small-outline packages (TSOPs), chip-scale packages (CSPs), and ball grid arrays (BGAs). Conventional molding compounds were used rather than new molding compounds designed for higher temperatures. TSOP leads were plated with either SnPb or lead-free NiPd. Spheres used for BGA or CSP components were either SnPb or lead-free SnAgCu. One-third of the components were reflowed at conventional reflow temperatures using SnPb paste; the other two-thirds were reflowed at the higher temperatures needed for lead-free SnAgCu paste.

Scanning and Results

Because the major concern was internal damage, all of the components were imaged by a scanning acoustic microscope at three points: before attach, after reflow, and after thermal cycling. Acoustic microimaging was used because it quickly provides both data and images on the internal features of a component and because it is nondestructive. It has another advantage as well: it uses a very-high-frequency ultrasound sensitive to internal features such as dies and lead fingers but is most sensitive to gaps, including delaminations, cracks, and voids likely to result from excess heat.

Figure 1. Acoustic image shows internal features of a TSOP in a JEDEC-style tray before placement (left) and after placement and reflow (right). Both solder paste and plating of leads used conventional lead-based solder. The image shows no delaminations, cracks, or other internal defects.

As expected, the study showed that components using lead-based solder for both paste and plating of the leads showed good stability—in other words, internal defects such as delaminations tended to be few and small. For example, one of the TSOPs using lead-based solder revealed a small crack in the molding compound before reflow, when the loose components were imaged in Joint Electron Device Engineering Council (JEDEC) trays. After reflow, this crack had hardly changed—a good demonstration of the inherent stability of the soon-to-be-replaced lead-based solders (see Figure 1).

TSOP components that used lead-free paste and tin-lead plating were also fairly stable. Some internal defects appeared after reflow, but most of these defects would be acceptable for most product reliability needs. However, these same defects would likely be unacceptable in medical applications.

Figure 2. A TSOP before placement (left) and after placement and reflow (right). Lead-free solder was used both for plating of leads and for paste. The higher reflow temperatures have caused internal delaminations (red areas).

TSOP components that used lead-free solder for both solder paste and plating of leads—and were reflowed at temperatures reaching 247°C—had the worst results. Most of these components had extensive internal delaminations that would be unacceptable in any application. One example showed extensive delaminations of the molding compound from the tie bar in a pattern that created a direct path to the exterior (see Figure 2).

In terms of performance and reliability, the defects seen in all of the component groups represent two distinctly different classes: defects that will cause electrical failures during the expected life of the component and defects that will not. Various standards exist for evaluating the threat posed by delaminations, voids, and other internal defects in components that use lead, but no analogous standards yet exist for lead-free components. The task facing assemblers is making lead-free components as stable and predictable as lead-based components have been.

Solutions

There are at least five steps that assemblers of medical electronics systems can take to make the transition easier, faster, and less costly:

1. When necessary, redesign (or ask a supplier to redesign) a component to meet the higher temperature requirements. The NEMI study suggests that the thermal pathways within a component strongly influence that component's ability to withstand the 260°C reflow heat without internal damage. Redesign might mean creating pathways by which heat can escape more rapidly from the component, or it might mean locating elements where they are less susceptible to heat—for example, placing more molding compound between a die paddle and the exterior.
2. Use acoustic microimaging to evaluate components after reflow. The high resolution of today's systems can quickly pinpoint problem areas, and saved images can be used to compare a component to earlier results. The planar acoustic images that illustrate this article are usually the first step in inspection and analysis, but cross-sectional and three-dimensional acoustic images can also delineate defects.
3. If they provide advantages, use the new higher-temperature molding compounds. But keep in mind that these compounds must perform two functions: They must provide moisture sensitivity to 260°C, and they must remain stable to the same temperature.
4. When necessary, look for a second source to supply particular components. There seems to be a strong willingness within the industry to share information for mutual benefit, but not all vendors will reach acceptable levels of component reliability at the same time.
5. To limit exposure to moisture, seal components in a dry atmosphere during early production, before attach. Typically, the four to five days of exposure used to lower tin-lead soldering temperatures will no longer apply.

Conclusion

Within a few years, systems using lead-free solder must reach the levels of reliability of conventional lead-based solder. The transition will have a price, but much of the cost will relate to labor and reliability rather than materials. Lead-free solders themselves will probably cost more than lead-based solders—NEMI's estimate is three to five times as much—but will have relatively little impact on overall costs because of the small quantities used. The impact will be greater, of course, in wave soldering, where larger quantities of solder are needed. Higher-temperature molding compounds may also be slightly more expensive than conventional varieties. As is the case for any manufacturing practice, careful planning and execution will be the keys to making the transition in a timely and efficient manner.

Tom Adams is a consultant to Sonoscan Inc. (Elk Grove Village, IL). He can be reached at info@sonoscan.com.

Copyright © 2002 Medical Electronics Manufacturing