By Dr. Malcolm Warwick, Loctite Multicore Corp., Hemel Hempstead, U.K.
Solder has been the most common, most reliable method of electronics interconnection since the birth of printed circuit boards, with SnPb solder leading the market. Neither solid metal solder nor solder paste changed dramatically until a few years ago, when concern surfaced about the use of lead in electronics assemblies. That concern, combined with the extraordinary advances in PC boards over the last decade or so, has put solder manufacturers in a position to reshape their industry. Board real estate is a major driver in the solder market today, second only to lead-free requirements. As circuit board components continue to shrink (Figure 1), leads are also becoming smaller, even as the number of leads per component is increasing.
Solder pastes perform a role in several aspects of the drive to miniaturization that are fusing IC packaging, hybrid micro-electronics and PC board assembly. Board-Level Applications At the board level, stencil printing of solder pastes has evolved to a point where printing onto pads for 0.4 mm pitch components is regarded as "normal," while, just a few years ago, major OEMs believed that this would be impractical in a normal production environment. Equipment, stencils and solder pastes have advanced so much that the best OEMs and CEMs operate very robust processes at this pitch. Array packages with small pads on the PC board and 0201 passive components are now either in planning or actually in production.
Most components self-align under the influence of the surface tension of molten solder. But if a passive component wets faster at one end than the other, it can be flipped up from the board before the second termination is soldered, which may cause a solder alignment defect know as "tombstoning," shown in Figure 2.
Board design and component solderability and placement are important in preventing tombstoning. When these measures fail, however, managing the wetting process during reflow is the only way to prevent the defect (Figure 3). Altering the reflow profile is a hit-and-miss activity, so solder paste manufacturers have adopted different strategies designed to slow the wetting process during reflow. While slower wetting is currently achieved through alloy selection, the future options for lead-free soldering are less clear. The need for finer-pitch printing has driven solder paste designers to employ powder with smaller particle sizes than previously specified. While most PC board assemblers use IPC type 3 powder, more are asking for type 4. Particle Size In reality, the particle size ranges quoted in specifications are somewhat arbitrary and it may not be necessary to make a step change from type 3 to type 4. Smaller solder powder can reduce the process windows for both printing and reflow, since the greater physical interaction between the powder and the flux vehicle changes the rheology. This change inevitably causes an increase in the quantity of metal oxide in the paste. Fortunately, paste manufacturers have the opportunity to build on experience gained from the use of very small size powder-particularly direct wafer bumping by printing and reflowing solder pastes through flip-chip assembly using pastes with type 6 solder powder. These processes demand consistency of print volume, which is related to stencil design and fabrication quality, as well as the carefully controlled rheology of the solder paste. At reflow, the use of an inert atmosphere-nitrogen in this case-results in the conditions needed to produce defect-free deposits and joints. Lead-free alloys, of course, are now also being used in these pastes. Meanwhile, solder pastes have been formulated to provide in situ bumping for small BGA packages. In this scenario, the challenge is to achieve consistent and relatively thick deposits that will provide sufficiently large solder balls after reflow.
Bridging Keeping the aperture of the stencil small so that adjacent deposits are farther apart means that thick stencils are required and paste release becomes a problem. Opening out the stencil apertures to reduce stencil thickness and improve paste release can lead to adjacent deposits that fuse during reflow to create a bridge (Figure 4). Applications Fine-pitch and lead-free soldering are both concerns primarily for portable consumer electronics applications. Manufacturers of military or business infrastructure products or other large, high-reliability applications have a different set of requirements.
Basic quality issues like good release characteristics and good printability remain important requirements where application components are large and miniaturization or fine-pitch printing is not an issue. No-clean pastes that are pin probe-able (for testing) are highly desired in these applications. A somewhat low-tech issue also causes problems for electronics manufacturers. Because the flux medium used to formulate solder paste is made from natural resin, a product of plants and trees, even identical resins can vary somewhat from one lot to the next, causing batch-to-batch inconsistencies in finished solder pastes. Solder manufacturers can address this problem through the development of modified natural products or purely synthetic resins.
Lead-Free Processes Any discussion of the future of the solder business would be incomplete without a look at current trends in lead-free processes and products. While both Europe and Japan have played an active role in the move to lead-free solders, the European WEEE (Waste from Electrical & Electronic Equipment) directive has gone through many recent changes and shifts, most recently encouraging recycling rather than requiring lead eradication. The elimination date for the designated hazardous materials, including lead, has also moved back to 2006 or 2007. Complete Elimination of Lead While there is no actual legislation pending, many Japanese manufacturers have taken a proactive stance by requiring the complete elimination of lead from their products. Several companies, including Hitachi, Matsushita, Sony and the cell phone division of Toshiba have established rigid deadlines to reduce or totally remove lead no later than 2002, according to the IPC Roadmap, Draft 4, (Figure 5). To be competitive in the Japanese and European markets, North American manufacturers clearly must begin to address these lead-free concerns. Pb-free soldering technology has been used in many industries, because of the desirable melting temperatures of these formulations (Figure 6). The electronics industry must use the Pb-free alloys, although they melt at a different temperature from the existing materials. Technical Consequences Broadly, the technical consequences of Pb-free soldering stem from the decision to go to a higher melting temperature for PC board assembly than tin/lead/(silver) eutectics, or to use alloys with lower melting temperatures. The majority of manufacturers have opted to live with higher melting temperatures since this is the only way to achieve adequate reliability in most service environments. Lead-free solder alloys increase melting temperature from about 180°C to about 220°C. While process temperatures do not need to move up by the full 40°C, physical damage to the board and component packages at elevated temperatures is an obvious concern with emphasis on the integrity of hermetic packages. Existing packages are being re-qualified, but there is a need for better materials that will withstand the rigors of temperatures as high as 260°C or even 280°C. In their efforts to reduce peak temperatures, process engineers will seek ways of minimizing the delta T across boards without negatively affecting solder paste fluxes.
Longer, Hotter Reflow Profiles Complex boards with a mix of components and relatively high thermal mass need longer, hotter reflow profiles to minimize both delta T and peak temperature, which destroys many fluxes that have been used for tin/lead soldering. Developments in both the fluxes and oven designs will be driven by the shift to higher reflow temperatures. For some board assemblers, the changes will be an incentive to switch on the nitrogen in the reflow oven to bring back a wider process window. Higher reflow assembly temperatures also affect soldered joints within components and modules. Alloys like tin/silver will melt during the Pb-free reflow process and may affect the reliability of the modules-depending on the initial integrity of the tin/silver soldered joints, the severity of the high-temperature excursion, and the nature of the bonding surfaces. Additionally, the surface tension of the molten tin/silver alloy in the module joints may not be sufficient to hold components in place and some kind of secondary fixing may be required. Modified Finishes Component manufacturers are now beginning to modify solderable finishes to make them compatible with Pb-free processes. While some of the finishes already on the market, those based on nickel and palladium, for example, are compatible with tin/lead and Pb-free processes, many are not. Clearly, the use of tin/lead finishes as metallization on lead frames or balls on BGAs has to be eliminated if the final assembly is to be Pb-free. Coatings of tin and tin/bismuth have received the most attention, although there is a lingering concern about the vulnerability of these finishes to whisker growth and even allotropic transformation (tin pest). The mechanisms for both defects are complex and may take a long time to appear.
Over the short term, PC board assemblers who are using bismuth coatings and tin/lead solders will face reduced solder joint reliability in higher temperature environments. The change in component finishes may precipitate an earlier change to Pb-free solders.
The evidence is building that the reliability of Pb-free soldered assemblies at least matches that of their tin/lead equivalents. However, published information has mostly related to "simpler" component geometries, with data only now appearing for BGAs and other components using Pb-free balls. Confidence in Pb-Free Solders Needed It will be a long time before engineers have the confidence to put Pb-free soldered joints into life-critical applications, as the learning curve for using tin/lead alloys in these applications was long and littered with errors. Improved mathematical modeling techniques may shorten the path, but there may still be a tendency to use secondary fixturing methods, such as underfills, in situations where they might not be strictly necessary.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2001