NEMI Report: A Single Lead-free Alloy Is Recommended
December 31, 1969 |Estimated reading time: 6 minutes
The National Electronics Manufacturing Initiative (NEMI) finds a consensus: the general-purpose lead-free solder alloy should be from the Sn/Ag/Cu family.
By Edwin Bradley, Carol Handwerker and John E. Sohn
There has been substantial discussion worldwide in the electronics assembly industry on the "ideal" composition for the primary lead-free solder alloy as a replacement for eutectic tin/lead (Sn/Pb):
- In Europe, the IDEALS project selected a range of solder compositions for processing printed circuit boards (PCB): Sn/3.4-4.1Ag/0.5-0.9Cu.1
- In May 2000, the NEMI Lead-free Assembly Project recommended the Sn/3.9Ag/0.6Cu (± 0.2 percent) alloy.2
- Major Japanese OEMs investigated numerous lead-free alternatives, including alloys containing bismuth and/or zinc, such as Sn/Ag/Bi/Cu, Sn/8Zn/3Bi and Sn/58Bi. However, over time, the Japanese industry has moved toward Sn/Ag/Cu alloys.
- The Japan Electronics and Information Technology Industries Association (JEITA) recently recommended the Sn/3.0Ag/0.5Cu alloy, partly due to concerns over patent issues.3 However, cross-licensing of nearly all the Sn/Ag/Cu family of solder alloys now appears complete, meaning all formulations are available from most solder suppliers. Accordingly, alloy selection should be driven by overall performance in product applications.
Following are some of the reasons for NEMI's alloy choice and variations of the system relative to melting point and fraction of melted material. Microstructure and reliability results also are discussed.
NEMI Selection Process
In approaching the lead-free solder issue, the NEMI team realized that they could make a major contribution to the industry by recommending a single solder solution to replace the Sn/Pb paste used for high-volume, surface mount component assembly. Since standardizing on a single alloy is particularly important to contract manufacturers as well as OEMs who seek to minimize their investments in equipment and facilities, it became the overriding goal of the project.
Figure 1. Solid phase fraction plot as a function of temperature for various Sn/Ag/Cu alloys. Charted are the NEMI alloy (blue), the JEITA (purple), the IDEALS (red), a very high tin (gold), and one high in silver and copper (green).
null
In making a decision, the NEMI team conducted a thorough literature and patent review and gathered all available data that the project's 30 participants, including five solder manufacturers, could contribute. The IDEALS project reports were particularly helpful in narrowing the alternatives considered.1,4 For example, the reports demonstrated that in through-hole joints during the cool-down phase, a solder with a large "pasty" range could lead to skewed components or delamination ("fillet lifting"). Thus, the solder manufacturers generally recommended choosing an alloy with no more than three elements for ease of solder manufacturing.
Based on the information reviewed, the team established the following criteria for selecting a solder alloy, which led to the choice of the Sn/Ag/Cu family of alloys as the most promising solution5:
- The melting point should be as close to Sn/Pb eutectic as possible.
- The alloy must be eutectic or very close to eutectic.
- It must contain no more than three elements.
- For ease of implementation, the use of existing patents, if possible, should be avoided.
- The potential for reliability should be equal to or better than Sn/Pb.
At the beginning of the NEMI project, the primary information on the family of selected alloys was the way the composition affected alloy melting points. Figure 1 shows the volume fraction of solid calculated as a function of temperature for different Sn/Ag/Cu alloys. The calculation is based on an earlier work by Moon and Kattner that conclusively identified the eutectic of the system as Sn/3.5Ag/0.9Cu.6
null
The melting points of the NEMI, JEITA and IDEALS alloys are within 7°C of each other. The NEMI alloy, Sn/3.9Ag/0.6Cu, is slightly more silver-rich than the eutectic, which will compensate for tin-rich platings that are likely to be introduced and that dilute the solder joint. The decision to go slightly higher than the eutectic with the silver composition was based on the fact that the amount of second phase that can form is very small (less than 1 percent) for temperatures over 218°C, and that its microstructure should be very repeatable as a result.
Microstructure
As manufacturers convert to lead-free processes, it is likely that components with Sn/Pb-coated leads and Sn/Pb ball grid array (BGA) solder bumps will continue to be used. Eventually, however, the industry will need to phase over to lead-free coatings. For this reason, the lead-free solder chosen also must be robust when some lead is present in the solder joint.
Microstructurally, the NEMI alloy has shown to be very consistent and uniform, both alone or when mixed with Sn/Pb (Figure 2). Intermetallic layer thickness at the copper and nickel substrates is nominally greater than that seen with Sn/Pb.
Reliability
The NEMI project focused on comparing the selected alloy with eutectic Sn/Pb for surface mount solder joints. The test plan permitted the team to compare lead-free solder with lead-free component coatings, with components featuring Sn/Pb coatings, and with both Sn/Pb solder and lead-free coatings. The table summarizes the results of the thermal cycling reliability testing on the various components.
The NEMI lead-free alloy consistently was as good as or better than the eutectic Sn/Pb alloy (to a 95 percent statistical confidence level). In many cases, the improvement was a factor of two or better. The mixed test cells, where the NEMI Sn/Ag/Cu solder paste was reflowed using typical Sn/Pb plated or bumped parts, generally performed as well as or better than the Sn/Pb alloy, with no catastrophic failures as had been seen with combining Sn/Pb plated or bumped parts with bismuth-containing alloys.7 Thus, there is little risk of major latent reliability issues if Sn/Pb parts were to make it into lead-free assemblies using the NEMI alloy.
Figure 2. CSP solder joint micrographs from a mixed cell (a) and a lead-free cell (b). Both are very uniform, with some Ag3Sn intermetallic plates visible in the lead-free solder joint.
null
null
Although the reliability testing was focused on comparing the NEMI alloy to Sn/Pb, one test cell was mixed using the NEMI BGA solder bumps assembled with the Sn/3.0Ag/0.5Cu solder paste. While it performed the same as the Sn/Pb cell statistically, performance was less than that of the lead-free cell containing both the NEMI alloy solder bumps and paste. More work is needed to confirm and understand that unexpected result.
Overall, much work has been performed on the Sn/Ag/Cu alloy over the past two years, both inside and outside of the NEMI project. The results have been uniformly good and point to the fact that the alloy meets the requirements of most, if not all, assembly applications. One common lead-free alloy is ideal to leverage solder manufacturing efficiencies and develop a baseline for modeling and reliability performance.
Conclusion
NEMI recommends a single lead-free alloy (Sn3.9/Ag0.6/Cu) to replace eutectic Sn/Pb solder for high-volume surface mount applications. The alloy has been used successfully in production and has demonstrated exceptional reliability in thermal cycling compared to eutectic Sn/Pb. Many more details about the NEMI alloy as well as the NEMI reliability testing will be detailed in a book to be published by IEEE Press/Wiley & Sons in 2003.
For a complete list of references contact the authors. Edwin Bradley may be contacted at Motorola; E-mail:edwin.bradley@motorola.com; Carol Handwerker may be contacted at NIST; E-mail: carol.handwerker@nist.gov; John E. Sohn may be contacted at NEMI; E-mail: jesohn@aol.com.