Issue link: https://iconnect007.uberflip.com/i/1185676
72 PCB007 MAGAZINE I NOVEMBER 2019 The one-way analysis graph shows that all the golds achieved the same pull strengths. The one ANOVA analysis of variance has been in- cluded to show that there exists a statistically supported variance. The significance is con- firmed by the Prb>F function of <0.0001. The scale of the diagram was, in this instance, too small for the statistical significance to have been borne out by the All Pairs Tukey-Kram- er analysis. According to this result, the pull strength result for CN Au 1 on the HP nickel was marginally better. In practical terms, Figure 11 acts as confir- mation that the cyanide-free gold bath (CNF Au) performs comparably to the cyanide-con- taining baths independently of whether the re- lated nickel diffusion barriers are HP nickel or MP nickel. As previously mentioned, the pull strength is only one aspect of wire bonding reliability. The other aspect is the fracture mode or break loca- tion. Essentially, only fracture modes 1 and 5 are of true concern and are considered failures. Fracture mode 3 represents the perfect result, while failures 2 and 4 may be indications that bonding parameters need to be optimized. In reality, fracture modes 2 and 4 are the most common as the welding process always entails a degree of wire thinning. It can be seen that in Figure 12 that the status quo for CNF is main- tained while the cyanide-containing immer- sion gold variations both show a small degree of wedge lift for the first bond. The minimal amount of level 1 fracture modes for CN Au1 and CN Au2 could be at- tributed to the wire bonding parameters, so in reality, the performance of all the immersion gold baths should be considered as equal. Al- though the results for the cyanide-containing gold baths were achieved by plating on dedi- cated electroless nickel baths, the results for the CNF Au were a conglomeration of results plated on HP nickel and MP nickel. To evaluate whether the electroless nickel has a defining impact, a one-way analysis was carried out, isolating the results for the CNF Au only. These results are shown in Figure 13. The pull test result for the MP nickel is sta- tistically superior to that for the HP nickel; however, the disparity is so low that it can be overlooked. The type of nickel has no veritable impact on the pull strength for the CNF Au. In Figure 14, the fracture modes are evaluat- ed. The results in Figures 13 and 14 contradict themselves if the differences were significant. The fracture modes for both nickel types can be regarded as the same. Solder Joint Reliability Solder joint reliability (SJR) is a measure of the formation of the intermetallic compound (IMC) and can be evaluated by physical shock Figure 12: Description of fracture mode used in this article. Figure 13: ANOVA, the pull strength for all immersion gold types in the study.