Abstract
Adhesion at the interface between dissimilar materials in the semiconductor industry is an important topic, but reliable quantitative methods for strongly adhesive or highly plastic layers are hardly available. This study aims to investigate the suitability of the cross-sectional nanoindentation (CSN) method for determination of the critical energy release rate of thin film stacks in the presence of a polyimide layer as a representative structure for such a case. For this purpose, the adhesion of a deliberately weakened Si/SiO(x) interface in a Si/SiO(x)/Al/Si(x)N(y)/polyimide stack is examined by systematic variation of the experimental parameters. This allows for a limitation of the plastic energy dissipated in the polyimide layer during delamination, while still investigating its influences on the overall delamination behavior. The results and evaluability of individual experiments were strongly affected by the geometry of the indenter tip and the distance of indentation to the interface, while other internal control parameters of the nanoindenter were more relevant for ease of the experimental procedure. An optimized choice of the mentioned parameters, 8 μm distance from the deposited layers for the cube-corner tip geometry, and 10 μm for the Berkovich geometry, led to 30 % of evaluable results, considered as high for quantitative adhesion testing of thin films. A multi-stage finite element analysis was developed to consider the effect of plasticity in the polyimide layer. Using this, the critical energy release rates ( Gc ) of the Si/SiO(x) interface was determined to be 9.61 J/m(2) and 10.85 J/m(2) for the cube-corner and Berkovich tip geometries, respectively. This work presents a novel promising way to extend the application fields of the CSN method for the determination of the energy release rate of systems containing layers with a high plasticity.