Effect of Surface Morphology on the Corrosion of Electroless Nickel/Immersion Gold Films

Electroless nickel with immersion gold (ENIG) process has been widely used as a surface finishing for PCB (Printed Circuit Board). The ENIG process provides a uniform coating on Cu layer without any complex patterning process, and so it has wide and diverse electronic industrial applications; fine pitch surface mount, ball grid array(BGA), flip chip BGA packages and so on. However, there is a reliability issue in the ENIG process. Many researchers have pointed out that severe corrosion of electroless Ni surface is the one of the major factors to reliability failure. In the extreme case, it is known as the black pad which caused a failure of solder joint. A content of phosphorous higher than Ni-P deposits was detected on the failed pad surface, and galvanic cell formation resulted into the corrosion due to the difference of P concentration. The Ni surface has a mud-cracked appearance and the phosphorous content is abnormally higher (exceeding 10 wt%). Another similar issue related to ENIG process is pad discoloration due to Ni oxidation on the top of ENIG pad. Such discoloration pad is regarded as a failure due to cosmetic reason and a lack of clear standards. Sometimes, it actually leads to high contact resistance due to Ni oxides on the top of ENIG layer. Excessive oxidation of Ni layer may result in the non-wetting of solder to the ENIG layer during the reflow process for solder ball attachment because it is difficult for the flux to remove such excessive oxides. The Ni oxidation may be owing to Ni cation diffusion passing through Au immersion during wet process such as rinsing steps in the downstream process after ENIG treatment, which is an electrochemical phenomenon as well. Here, the pad discoloration of Ni corrosion during Electroless nickel with immersion gold (ENIG) treatment was investigated by SEM, XPS and cross-sectional EDS analysis. The Ni corrosion happened at the nodule boundary for relatively thinner layer whereas the corrosion pits were generated in the thick Ni deposits. The electroless Ni was amorphous structure, which meant that the domain boundary energy between two Ni nodules did not change during Ni growth. However, the cusp stress at the surface intersection of two nodules gradually decreased during Ni growth, and hence the corrosion mode changed from nodule boundary corrosion to pit corrosion. It can be surmised that the nodule boundary corrosion can be suppressed by increasing Ni nucleation as well as Ni thickness.