Optimizing the wire-bonding parameters for second bonds in ball grid array packages
Journal
Journal of the Chinese Institute of Engineers, Transactions of the Chinese Institute of Engineers, Series A/Chung-kuo Kung Ch'eng Hsuch K'an
Journal Volume
23
Journal Issue
5
Pages
625-632
Date Issued
2000
Author(s)
Abstract
Abstract The bonding parameters for the second bonds in BGA were optimized using the Taguchi L27 experimental design. The bond wire was a 33 μm gold wire, and the bond pads had Au/Ni/Cu three‐layer structures. The objective function for the optimization was the pull strength. In the pull test, the pull hook was placed at a location closer to the second bond than to the ball bond. In such a testing configuration, all failures occurred at or near the second bonds. At the bond temperature of 150 °C, it was found that the optimal ultrasonic power was 0.25 W, the optimal bond force was 0.4 N, and the optimal bond time was 0.6 s. The average pull strength at the optimal conditions was 0.126 N. These parameters also produced bonds that had the best signal‐to‐noise ratio in pull strength, suggesting that the consistency in pull strength is also the best. From ANOVA techniques, it was found that power, time, and the interaction of power and time had the strongest effects on the pull strength and the signal‐to‐noise ratio. Through SEM observation, three fracture modes were identified: the interface fracture mode, the heel fracture mode, and the mixed mode. The overall strength of the bond is mainly determined by two contradictory factors, the strength of the interface between the bond wire and the pad, and the heel strength. The optimal bonding parameters represent a balance between these two factors.
Type
journal article