Mechanical fatigue properties of heavy aluminium wire bonds for power applications

In this study, a technology-oriented simplified mechanical fatigue testing approach for aluminium heavy wire bonds as well as first experimental results are presented. In the test setup, bonding wires were displacement-controlled loaded with different amplitudes at room temperature and the corresponding cycles to failure were experimentally determined. Loop geometries were varied in a technological meaningful range. The experimentally determined endurance curves show a strong influence of the bonding geometry on the lifetime of the bonding wires. In addition to testing, a three dimensional finite element model of the different bonding wire geometries was developed in order to quantify the local deformation situation at the failure site in terms of equivalent strain. The global mechanical p roperties used for the simulations were determined by tensile tests of unprocessed bonding wires. The experimental results in terms of number of cycles to failure could be represented as a function of the change in equivalent strain at the heel for the different bond loop geometries. Using a common double-logarithmic endurance plot, the results for the different bond loop geometries could be approximated by a linear dependency. This result is in accordance with the expectation that a Coffin-Manson approach can be applied to predict the life time of the aluminium wire bonds. From these results, it can be concluded that the experimental testing approach and the applied simulation model is applicable to understand the effect of different bonding loop geometries on the number of cycle to failu re. For a more generalized understanding, it has to be taken into consideration that the mechanical properties close to the heel were affected by the bonding process prior to fatigue loading. In form of a preliminary study, it is shown that spherical indentation testing on cross sections of the bonded wires provides a useful methodical approach to characterize these variations and to extract the local material properties for further expanded modelling