Increasing the robustness for reliable packages by prediction of delamination by cohesive zone element simulation

Robustness of a package is often proven by performing temperature cycling tests. Thermo-mechanical stress caused by the mismatch of coefficients of thermal expansion (CTE) and temperature variations remains a major concern for the reliability of semiconductor components. This issue is usually addressed by exposing the component to a certain number of cycles, followed by e.g. scanning acoustic microscopy (SAM) to investigate delamination. Discussions about specific cycling conditions, e.g. using -65°C/+175°C instead of -55°C/+150°C for the minimum and maximum temperatures of the cycles or even using liquid-liquid cycling instead of air to air to speed up investigations [1], are often moot, because no real understanding of the effect of the cycling conditions on the component is available. Introducing the parameter adhesion allows the prediction of the delamination behaviour and the possibility to identify critical locations. Furthermore, it seems to be feasible to distinguish between two different failure mechanisms (interface crack propagation and crack initiation) leading to the same failure mode (delamination). Furthermore, it is almost a truism that testing alone does not suffice to ensure the reliability of a component. Reliability has to be built into the components from the beginning. As a consequence, the question should be turned around: It is not enough to look at delamination after a certain number of cycles in a stress test, the questions rather are, how the component should be designed and how the materials should be chosen to prevent delamination. Thus, the focus is changed from measuring delamination to measuring adhesion.