Sn-based solder design using machine learning approach

Sn-based solder is a typical material for electronic interconnections. Low-Ag-content Sn-based solder exhibits a low hardness and is promising for avoiding the drop-induced failure. However, Sn-based solder has a wide range of composition and thus modulating its hardness via experimental trial-and-error method is not economically feasible. In this study, we employed the machine learning approach to design low-hardness Sn-based solders. We successfully designed an alloy that showed a hardness value lower than pure Sn, which has the potential of having a high-resistance to drop-induced failure.