In situ aging study on the variation of Sn0.7Cu/Cu solid interface marked by bubbles

Abstract

As the joint reliability of lead-free soldering is an important issue in electronic packaging industry, the research of interfacial bubbles is extremely imperative due to their close relationship with the quality of solder joints. In the process of solid state aging: the shape, size and location of bubbles will remain almost unchanged, subsequently allowing them to be used as markers during the procedure of our experiment. Based on this, the motion of Sn0.7Cu/Cu solid interface and the growth characteristics of the interfacial intermetallic compounds (IMCs) were in situ investigated in this study, using the scanning electron microscope (SEM) for the cross section each time after aging for a period of time at 150°C from 0h to 500h in oil (the soldering reaction part was carried out in furnace with the soldering temperature of 250°C and holding time of 60s, followed by water cooling). It was found that the solid interface would move towards the direction of copper substrate relative to the reference bubble, with its speed declining as the time passes. As the average composition of IMC layer in the normal area far from the bubble had an overall approach to the stable final substances Cu6Sn5 and Cu3Sn, the moving phenomenon can be explained by the transformation on the IMC/Cu interface including a volume change. On the other hand, the outline of the interfacial grains would initially have distinct appearance. On the perfect Cu6Sn5/Cu interface, a layer of Cu3Sn appeared at the beginning of the aging procedure, and its growth orientation was obviously inclined outward and perpendicular to the cross-section surface, and in contrary to this, the region under the bubble didn't generate a Cu3Sn layer. An increase in volume during a chemical reaction would produce some form of deformations through compressive stress, whereas, a decrease in volume would create a tensile stress which will eventually cause some flaws and voids on the reaction interface; maybe these two different forms of transformation can be distinguished by experimental microcosmic observation in situ or from calculation of the scope of the correlation between mole volume and temperature for specified materials. At 150°C the diffusion rate of copper in Cu6Sn5 layer is larger than that in Cu3Sn layer as put forward by some researchers, and, thus, the existence of the intermediate Cu3Sn layer will block the spreading of copper. Most probably, the volume changes caused by the appearance of the new product layer and the special diffusion feature at the aging temperature can be the highlighted parameters for the phenomena mentioned above.