Electroless nickel / immersion gold (ENIG) has been recognized within the electronics packaging industry as a reliable surface finish for soldering applications. Unlike surface finishes, such as OSP, immersion silver or immersion tin, which produce a Cu-Sn intermetallic upon soldering, the nickel layer of ENIG acts as a barrier to minimize copper dissolution during the soldering application. This function is especially important for assemblies that require multiple reflow steps. However, if not properly controlled, excessive corrosion of the nickel by the immersion gold step can also cause the well documented “black pad” effect, often resulting in solder joint reliability issues, including brittle fracture. As an alternative to ENIG, an electroless pure palladium layer over the electroless nickel (ENEP) provides a solderable surface without the noted risk of damage to the underling nickel layer. By eliminating the corrosive attack of the gold bath, the final layer stack would be suitable for high reliability soldering involving both IC substrate and PWB applications. In addition, because of the current high price of gold, replacing it with a pure palladium deposit can offer measureable cost savings for the ENEP process in comparison to ENIG. The paper summarizes an in-depth study of the effect on solder joint reliability caused by replacing the immersion gold by an electroless pure palladium layer. The study includes evaluations for both Pb-free (Sn-3.0Ag-0.5Cu) and eutectic SnPb (63Sn-37Pb) soldering applications. Results of investigations include: (1) cold ball pull testing to evaluate solder joint integrity, (2) SEM examinations of the underlying nickel surface, (3) IMC examinations to quantify nickel thickness degradation after multiple solder reflow cycles and (4) surface wetting through solder spread examinations. The paper discusses the relatively simple surface finish that, if proven effective in large-scale fabrication, may offer measureable performance and cost benefits in comparison to the more traditional ENIG finish.
Read full abstract