BGAs with lead-free solder balls are beginning to be shipped to assembly operations that are using eutectic Sn-Pb solder paste for surface mounting. What are the critical considerations?

Most lead-free BGAs are shipping with Sn-Ag-Cu (“SAC”) solder balls. The ternary eutectic composition is approximately Sn-3.5Ag-0.9Cu (wt. %), with a melting point of 217°C. The solder ball composition is commonly Sn-3.5Ag-0.75Cu, or Sn-3.8Ag-0.7Cu (wt. %), or a very similar chemistry. Presently, it may be possible to find lead-free BGAs with different SAC chemistries, Sn-3.5Ag eutectic or Sn-0.7Cu, among others, so it is critical to know the BGA ball chemistry. When using Sn-Pb eutectic solder paste to surface mount a lead-free BGA with typical SAC solder balls, it is important that the BGA balls melt during the reflow process. The eutectic Sn-Pb solder will melt at approximately 183°C, and then begin dissolving the SAC ball as the furnace temperature rises. If the ball does not melt, upon cool down it is likely that the interface between the SAC ball and the frozen eutectic Sn-Pb, modified by partial dissolution of the SAC ball, will contain voids. Making sure that the SAC balls are all heated to an approximate minimum temperature of 225° to 230°C, since the typical solder ball composition is not the SAC eutectic, will ensure the BGA balls will melt and mix with the Sn-Pb eutectic solder, greatly increasing the probability of a void-free interface. Another major consideration is making sure that the PCB and other components on the board can tolerate the reflow temperatures that may be needed to ensure melting of the SAC solder balls.

Why are some companies concerned about tin whiskers from matte tin-plated components? Isn’t it correct that during reflow soldering, the whiskers will dissolve in the lead-free solder paste, eliminating a pure tin surface?

The matte tin (Sn) finishes and annealing processes used today greatly reduce the risk of whiskers from matte tin. It is true that one can expect the lead-free solder paste to dissolve all the matte Sn from the pads of a “leadless” leadframe package, like a QFN. However, it is unlikely that the surface mount solder paste will wash away all the matte tin on the surface of gull-wing or “J” lead devices, such as a QFP or a PLCC. During reflow of QFPs and PLCCs, the matte Sn on the lead surface will melt at 232°C if all the leads are exposed to this temperature. Meanwhile, the lead-free or eutectic Sn-Pb solder paste will have melted and wetted the PCB pad and lead surface. Typically, surface tension of the molten solder paste will prevent the solder from wetting up the long QFP and PLCC leads. Consequently, upon cool down, it is quite likely that reflowed matte tin will be frozen on some lead regions, especially near the mold compound where the lead disappears into the package body. While much of the Sn may have been consumed in these areas because of the formation of intermetallic phases, it is likely that some pure tin will remain on the surface because the matte Sn specification typically demands the deposition of a minimum of 7 micrometers of matte Sn.

Correction to previous Q&A on the use of Matte Sn over Ni-plated Cu leadframes: The European semiconductor companies referred to in the previous note reversed an earlier position supporting the use of matte Sn directly on the Cu surface or matte Sn over a Ni- plated layer over the Cu surface. Work in the past year has indicated a risk of cracking the Ni layer during lead formation. Consequently the use of the Ni layer is no longer permitted.