Issue link: https://iconnect007.uberflip.com/i/992528
JUNE 2018 I PCB007 MAGAZINE 65 thin and uniform compared with that in case of IGEPIG and EPIG. It seemed that this thin and uniform IMC resulted in excellent SJR. The difference of SJR for P contents in EP lay- er for ENEPIG deposits had already reported [5) . The cross-section images of deposits before mounting solder ball were also observed as shown in Figure 5 to consider the difference of EPIG and IGEPIG. A lot of voids were ob- served between Pd layer and Cu for EPIG de- posit. On the other hand, it is difficult to con- firm any voids in this magnification for IGE- PIG deposit. Generally, voiding between cop- per and plated metal is influenced by copper surface roughness, the type of electrolytic cop- per, the kind of palladium activator, soft etch- ing, etc. It was assumed that the palladium ac- tivator process caused these voids because this process is a displacement reaction with cop- per and these voids had some worse influence for SJR of EPIG deposit, compared with that of IGEPIG deposit. Wire Bonding Reliability WBR was evaluated by pull strength of gold wire pull test for the sample as plated and after heat treatment as shown in Table 7. For the sample as plated, wire pull strength of EPIG, ENEPIG and IGEPIG de- posits were all sufficient. For the sample after heat treat- ment (HT) for 16 hours at 175°C, wire pull strengths of EPIG and ENEPIG deposits decreased. On the other hand, though the strength for IGEPIG deposit became slight- ly worse after HT, the value was still sufficient and comparable to ENEPIG and EPIG deposits without heat treatment (the sample as plated). The wide scans of AES were measured at the top Au surface of each deposit after HT as shown in Figure 6 to understand the difference of pull strength of Au wire. Though Pd for EN- EPIG, Cu and Pd for EPIG were detected, these metals were not detected for IGEPIG deposit. The ratio of Pd, Cu, Ni, and Au intensity at Au surface after HT for each deposit was calculat- ed from AES results as shown in Figure 7. Ob- viously, IGEPIG deposit prevented Cu and Pd diffusion. The comparison between EPIG and ENEPIG showed EN barrier prevent Cu diffu- sion. Wire bonding reliability related with the Au ratio at the top surface of each final fin- ishing. Because Au ratio of IGEPIG was quite high even if it was after heat treatment, IGE- PIG deposit had better potential of WBR than Table 7: Pull strength of gold wire for each deposit, before and after heat treatment for 16 hours at 175°C. Figure 5: Cross-section image by FIB-SEM; (a)EPIG, (b) IGEPIG; *Ni; protecting layer of Au layer during FIB.