The new 'mixed technology' PCB: what will you do when SnPb balls are discontinued?

Remember when the term "mixed technology" referred only to a design that included both surface mount and through-hole components? Ah, the good old days of SnPb. These days, mixed technology often refers to mixed alloy systems between components and solders, or "mixed metals technology."

The most common mixed metals combination is SAC 305 or 405 BGA balls in a SnPb soldering process. Every day it seems another manufacturer discovers the BGA package they've used for years will change from SnPb to SAC alloy, and the packages with SnPb balls will be discontinued. This puts assemblers in a quandary, as the specific component needed to meet production commitments is no longer available.

The first response is short-term containment: Contact all distributors of this component and try to procure enough stock to meet the demand forecast for a particular time frame. The next step is to look to the long-term corrective action: the course of action when the SnPb packages run out.

Reballing is one option, but it can be expensive and time-consuming, with numerous potential pitfalls and little published long-term reliability data for IPC Class 3 and other high-performance or exempt electronics. The reballing process requires two additional thermal excursions of the device: one to remove the Pb-free balls and their intermetallic compounds from the (usually ENIG) substrate pad, and another to attach SnPb balls. Proper MSD handling and thermal conditioning protocols absolutely should be followed during such a procedure. Defects that can be induced during reballing, such as popcorning, warpage or localized delamination, might not be captured during the process and can result in failures during assembly--or worse, in the field. If the BGA provider is willing to supply the package without balls, which often negates any "warranty of merchantability" liability claims, the assembler can procure the component prior to its first thermal excursion of attaching the Pb-free balls and subsequently avoid the thermal excursion inherent with their removal. Thus, the part reaches the assembly line with only one reflow cycle: ball attach (most likely with a SnPb alloy).

If reballing is the only option, the assembler has a choice: Do the work in-house, or contract it to a specialist. Keeping operations in-house provides greater control over supply-chain logistics, but requires considerable resource and financial investments. Outsourcing usually means it is administered by an organization with a core competency in the process. These organizations employ best practices to avoid potential pitfalls and in many cases may be the safest bet for long-term reliability or for short runs of SnPb assemblies prior to Pb-free transition. The cost-benefit model of whether to reball in-house or out must be constructed on a case-by-case basis, as the figures vary with each product, operation and transition time frame.

Running Pb-free balls in a SnPb process is becoming an increasingly popular option. Many reliability studies are ongoing, and complete results have not been published. A primary concern appears to be getting full mixing of the lead into the BGA ball. Although we may find that full mixing is not necessary, most assemblers are erring on the side of caution. Some want to achieve ball collapse. To get the ball to collapse, the device interconnects (bumps or balls) must reach peak temperatures above their melting points, roughly 221[degrees]C.

A recent iNEMI study showed that full mixing can occur without ball collapse. Because metals mixing is a function of time and temperature, these two factors were key variables in the experiment. One of the key findings was that temperature has a greater effect on mixing than time above SnPb liquidus (183[degrees]C). Assemblies that peaked at 215[degrees]C for 60 sec. showed a greater degree of mixing than assemblies that peaked at 210[degrees]C for 120 sec., especially for the larger ball volumes associated with coarser pitches.

So we now have solid data that establish higher peak temperature is more effective than TAL at achieving mixing; and we know that if we want the balls to collapse, we need to get even hotter. We've cut the SnPb peak temperature window by half to aid mixing, and by three-fourths if we want collapse.

Assuming we feel comfortable taking laminate and other components to the high end of the SnPb reflow window, can our current SnPb solder paste still do the job? Maybe, maybe not. Factors in this equation include thermal mass of the PWB, thermal transfer efficiency of the reflow oven, and the chemical constituency of the paste flux.

This is where the hybrid solder pastes I mentioned last month come into play. What we refer to as hybrids--pastes made by blending SnPb powder with Pb-free flux--have certain advantages, but should be approached with caution. The advantages include much better thermal stability in the high end of the SnPb reflow window, and formulations specifically designed to promote wetting to the SAC alloys from which BGA balls or bumps are formed. Numerous solder paste providers now offer this hybrid class of products.

Hybrid paste risks include lower electrical reliability and increased voiding rates. A Pb-free flux designed to run in a hotter reflow process may not provide desired electrical reliability when run in SnPb. The paste should be subjected to surface insulation resistance or electrochemical migration tests when processed under the SnPb profile, before it is introduced into production. The same goes for voiding. A flux designed for Pb-free processing can volatilize until 217[degrees]C when used with SAC alloy solders, but its outgassing paths will be closed at 183[degrees]C if used with SnPb solders. First-and second-generation Pb-free paste fluxes exhibited serious voiding when combined with SnPb solder. Modern Pb-free products appear more robust against voiding at lower process temperatures.

Finally, if using intrusive reflow processes, examine the ability of paste overprints to pull back to the PTHs. We know that different pastes have different "sweet spots" in the reflow window with respect to pull back; users should ensure the hybrid paste selected will support pin-in-paste processes without leaving random solderballs on the board.

This month's lesson is not one we have learned, but one we are still in the process of learning. As hybrid paste use becomes more common, industry will understand more nuances of its application. The upside is that it seems we do have a few viable solutions to the impending extinction of SnPb BGA packages. The downside is we need to proceed with caution regardless of the option we exercise. But it won't stop there. Before we get this solution nailed down, we need to start on a solution for the next-generation of mixed metal systems: BGA balls of low silver SAC alloys (SAC 105, SACX, or a more exotic low-silver alloy with special dopants) combined with SAC 305 pastes. Last month I said I love surface mount because it never gets boring. The mixed metals systems are a perfect example of the never-ending predicaments we continue to resolve to make electronics smaller, lighter and more powerful.

Chrys Shea is an R & D applications engineering manager at Cookson Electronics (cooksonelectronics.com); [e-mail]. Her column appears monthly.

[ILLUSTRATION OMITTED]