Abstract
Rapid growth of intermetallic compounds (IMCs) and the formation of Kirkendall voids during solid reaction are critical issues for reliability of Cu-based solder joints. Adding minor 4th element into solder joints is proposed to improve the mechanical reliability of joints. Pd is another element of interest due to higher mechanical reliability of joints with Electroless Nickel/Electroless Palladium/Immersion Gold (ENEPIG) as compared to that with Electroless Nickel/Immersion Gold (ENIG). However, for Cu-based solder joint, the Pd distribution and detailed mechanism how Pd influences the interfacial reaction and reliability of joints are not yet proposed. In this study, Sn3.0Ag0.5Cu (SAC305) solder doped with 0~0.5 wt.% Pd was used to reflow with Cu pad. For liquid-state reaction, Pd tended to dissolve into interfacial Cu6Sn5 and refine the grain size of Cu6Sn5. After multiple reflows, the IMC growth rate was enhanced by Pd addition due to the higher number of diffusion channels between Cu6Sn5 scallops. Nevertheless, the impact test showed that the Pd doping would increase the bonding strength of the Pd-doped joints. In solid-state reaction, the growth of Cu3Sn and Kirkendall voids were suppressed by the Pd doping into joints. X-ray elemental mapping and quantitative analysis demonstrated that Pd atoms were accumulated at the lower part of (Cu,Pd)6Sn5 layer. Based on thermodynamic calculation, the Gibbs free energy of (Cu,Pd)6Sn5 is more negative than that of Cu6Sn5, indicating that (Cu,Pd)6Sn5 is thermodynamically stable than Cu6Sn5. Therefore, the Pd-enriched (Cu,Pd)6Sn5 layer is regarded as a barrier to suppress the growth of Cu3Sn and the formation of Kirkendall voids. The detailed mechanism of Pd influence on interfacial reaction in solder joints was probed and discussed. In addition, the impact test was also employed to evaluate the Pd effect on the mechanical reliability. By the suppression of the Cu3Sn growth and the prevention of Kirkendall voids formation, the Pd-doped joints exhibits higher impact force than non-doped joints. The crack propagation via interfacial IMC indicated that the mechanical properties of Cu6Sn5 and (Cu,Pd)6Sn5 might dominate the performance of solder joints in impact testing. Furthermore, the indentation data showed that (Cu,Pd)6Sn5 exhibited higher fracture toughness and lower Young’s modulus than Cu6Sn5. It is argued that the difference in mechanical properties between Cu6Sn5 and (Cu,Pd)6Sn5 was believed to be the main cause of higher impact reliability in the Pd-doped solder joints. To amplify the advantage of the Pd doping during soldering with reducing amount of Pd addition into solder balls, the Pd-doped solder-on-pad (SAC-Pd SOP) surface finish was developed. The primary Cu6Sn5 and (Cu,Pd)6Sn5 formed at the interface of SAC SOP and SAC-Pd SOP system after 1st reflow, respectively. With the attachment of SAC305 solders during 2nd reflow, the IMC thickness decreased in SAC/SAC SOP joints owing to the increasing volume of solder matrix. However, the IMC thickness increased in the Pd-contained joints, resulting from thinner initial intermetallic thickness and stabilization of Pd-dissolved Cu6Sn5. Nevertheless, the IMC thickness in the Pd-contained joints was smaller than that in SAC/SAC SOP joints. In addition, the correlation between microstructure and related impact reliability in the SAC/SAC-xPd SOP joints was addressed and proposed. Finally, this study aimed to evaluate the potential application of novel Pd-doped lead-free solders for future solder designs.