Fracture characterislics at the interface of ultrasonic bonds between Au and AI were characterized by SEM following pull lesting to effccl separation of the bonded joints. Vertical seclions at Ibe bonding point weir produced by ion-sputter thinning, alld were examined by TEM. Results show Ihat the Ibiekness of the Au/AI atomic diffusion intcrfacc was about 5O0mn due to combined effccts of ultrasonic and thennal energy. Ultrasonic vibration activates dislocations in tile crystallinc latliee and increases alomic diffusion. The fracture morphology on the lift off inlerface was dimpled rapture. Tensile fracture occurred during lbe pull test not al the bonded interface but in die base maleftal: the bond strength at the interlace was enbanced by the diffnsion reactions that occu~ed across the intert~cc due to tile combined ultrasonic and thermal energy

The authors demonstrate that the ultrasonic vibration in flip chip FC bonding results in the generation of dislocations, and the atomic diffusion can be activated more easily along the dislocation lines which perform the fast diffusion channels, thus the dislocation diffusion is probably more prominent than the body diffusion during ultrasonic bonding. To minimize the intermetallic compound layer, the effectiveness of a different bonding approach is confirmed. Furthermore, an experiment-based mode of ultrasonic energy conversion was found that the ratio of up interface to down interface in ultrasonic FC bonding was about 2.3:1.

This study seeks to quantify the screw-fastening effect of the tool/transducer on wire bonding performance. Aluminum wire bonding experiments were performed on a laboratory test bench. Characterizations of bonding process through a wavelet analysis were used to study the relation between screw fastening, vibration behaviors and bond strength. The time-frequency plots were depicted for identifying un-modeled wire bonding dynamics. Finally some statistically time domain features were presented for further analysis.

This paper presents a recent study by monitoring input power in wire bonding process on its performance. The instantaneous driving voltage and current to the PZT/transducer system were recorded and the input power histories for all tests were analyzed. A stable and satisfied bonding can be obtained at moderate ultrasonic power setting. A laser Doppler vibrometer was used to record the response of the structure. The initial power oscillating may represent the phase locking chaos, and the final attenuation may reflect the remains of kinetic energy in the structure. Strength of wire bonding should be attributed to the input power during the main loading segment.