WELDING RESEARCH A B Fig. 7 — The thickness (X) of PdSn4 and Ni3Sn4 intermetallic compounds formed during diffusion soldering of a Pd/Ni couple at various temperatures NOVEMBER 2016 / WELDING JOURNAL 445-s microscopy using an accelerating voltage of 15 kV, and their chemical compositions were analyzed with energy dispersive x-ray spectroscopy (EDX). The lateral resolution of EDX analysis was about 1 m. The shear strengths of the various Pd/Ni assemblies were tested with a DAGE 4000 bond tester according to the IPC-TM650 (TM2.4.42.2 for die shear strength test). For the testing, the shear height and shear speed were set at 0.1 mm and 0.3 mm/s, respectively. Results and Discussion Figure 4 shows the microstructure of Pd/Ni joints in Case I of Fig. 3 after diffusion soldering at 275º to 350ºC for 30 min. In this case, the bonding between the Pd sheet and Ni plate at 275ºC for 30 min failed due to insufficient reaction at the Pd/Sn/Ni interfaces. For diffusion soldering at temperatures between 300º and 350ºC, it can be seen that the Pd sheet and Ni plate reacted with the Sn interlayer to form an intermetallic layer with a composition (at-%) of Pd : Sn = 52.9 : 47.1, which could be identified as the PdSn intermetallic phase in the Pd-Sn phase diagram (Ref. 14). On the other hand, the Ni plate reacted with the Sn thin film to form a reaction layer, which possessed a composition (at-%) of Ni : Sn = 43.1 : 56.9. The composition is near the Ni3Sn4 intermetallic phase in the Ni-Sn phase diagram (Ref. 15). The thicknesses (X) of the PdSn intermetallic layers were much greater than those of Ni3Sn4, and those of both intermetallics increased with the bonding temperatures — Fig. 5A. The apparent activation energies (Q) for the growth of both intermetallics can be estimated from the slopes of the Arrhenius plots in Fig. 5B for these intermetallic thicknesses (Ln X) vs. reciprocal bonding temperatures (1/T) that can be calculated as shown in Equation 1: X = A0exp(–Q/RT) (1) The results indicated that the apparent activation energy for PdSn intermetallics growth (40.6 kJ/mol) is much higher than that for Ni3Sn4 (26.8 kJ/mol), which implied that the interfacial reaction for the growth of PdSn is more sensitive to the bonding temperature than that for the Ni3Sn4 intermetallic compound. Furthermore, Fig. 4 also reveals that the Sn interlayer was completely exhausted under this condition, which prevented the melting of Pd/Ni joints during the hydrogen purification process at temperatures over 350ºC. However, long cracks appeared between the Ni3Sn4 and PdSn intermetallic layers, and the bonding strength was very low. The appearance of cracks at the Ni3Sn4/ PdSn interface is similar to that reported by Bader et al. for diffusion soldering of Ni/Ni and Cu/Cu couples using Sn interlayers (Ref. 16). In a study by Chang et al. (Ref. 17), it was verified that the interfacial voids that form between Ni-Sn or Cu- Sn intermetallic compounds can be eliminated by inserting an Ag3Sn intermetallic layer between the diffusion soldered Ni/Ni or Cu/Cu couple with a Sn interlayer. The diffusion soldering process in this study was modified to include a Sn/Ag/Sn multilayer between the Pd sheet and Ni plate (Case II in Fig. 3). for 30 min using Sn/Ag/Sn interlayers (Case II). Fig. 8 — Shear strengths of the SLID bonded Pd/Ni couple using Sn/Ag/Sn interlayers.
Welding Journal | November 2016
To see the actual publication please follow the link above