Adhesion of WTi to polyimide measured by complementary methods

Quantification of thin film adhesion to their substrates is crucial to ascertain reliable multi-material components, which are typically encountered in technologically relevant structures such as microelectronic devices. Metal-polymer interfaces have gained increased attention in the past since they open a new research field towards developing flexible electronic devices, yet their quantification is often experimentally challenging since standard routes to assess adhesion such as nanoindentation-based techniques or four-point bending may not be applicable due to lack of rigidity. In the present study interface adhesion of nominally 150 nm thin WTi films sputtered onto polyimide, part of a multi-material stack consisting of brittle and compliant materials, was determined experimentally. Upon contactless removal of the bilayer by etching a sacrificial layer, originally part of the multilayer stack, the WTi film showed cracks and buckles, which could be characterized by atomic force and confocal laser scanning microscopy. Hence, two models to quantify adhesion were compared: the model of tensile induced delamination – typically employed for metal films on compliant substrates – and the model of Hutchinson and Suo, which is usually applied on spontaneously formed, straight buckles. The first model reveals adhesion values for the native WTi-polyimide interface of 2.4 ± 0.8 J/m2 and the latter model 4.7 ± 2.3 J/m2. The slight difference in adhesion results, which are both rooted in the same theory, could be rationalized by the fact that the model for tensile induced delamination gives a measure of the lower bound of the adhesion energy. The overall rather lower adhesion values could be explained by additional transmission electron microscopy investigations conducted on the cross section of the interface, which reveals no interlinking morphology and a flat interface.