An in-situ experimental-numerical approach for interface delamination characterization

Interfacial delamination is a key reliability challenge in composites and microelectronic systems due to (high density) integration of dissimilar materials. Delamination occurs due to significant stresses generated at the interfaces, for instance, caused by thermal cycling due to the mismatch in thermal expansion coefficient and Poisson’s ratio of the adherent layers. Predictive finite element models are generally used to minimize delamination failures during the design and optimization of these materials and systems. Successful prediction, however, requires a relevant interface model that can capture the observed (irreversible) crack initiation and propagation behavior in experiments. To this end, dedicated delamination experiments with in-situ microscopic visualization are needed to identify the relevant delamination mechanism(s) and to accurately measure the interface properties, such as the interface toughness, as a function of mode mixity (i.e. loading angle). Hence, the goal of this research is to develop experimental-numerical tools required for accurate characterization and prediction of interface delamination. As a first step to reach this goal, a novel Miniature Mixed Mode Bending (MMMB) delamination setup, which enables in-situ characterization of interface delamination in miniature multi-layer structures, was designed and realized. This setup employs an inventive loading configuration to sensitively measure global load-displacement delamination curves for the full range of mode mixities from which the interface toughness or Critical Energy Release Rate (CERR) can be determined, while it was designed with sufficiently small dimensions to fit in the chamber of a scanning electron microscope or under an optical microscope for detailed real-time fracture analysis during delamination. The performance of the setup was assessed using dedicated test samples, supported by finite element analyses. The measurement concept was successfully validated on homogeneous bilayer sampleswith a glue interface system. The validation experiments also revealed roomfor improvement of themeasurement accuracy, robustness, and applicability. Therefore, further optimization in the design was performed and an improved version of the MMMB setup was developed. This setup can access a considerably larger range of interface systems, shows significantly higher accuracy and reproducibility in load-displacement measurements, and is more robust. The potential of the new in-situ experimental technique for interface parameter identification was also illustrated. For instance, high resolution in-situ SEM imaging during delamination allows for measurement of the strain maps and crack opening displacement (COD) fields using digital image correlation in addition to the identification of the delamination failure mechanism. In-situ SEM observation of delamination in different interface structures reveals failure mechanisms ranging from interface damage to interface plasticity. Hence, an irreversible model description of the interface behavior that can capture the observed unloading-reloading responses is needed for accurate prediction of, for instance, crack branching and crack propagation at multiple interfaces using predictive finite element models. Therefore, a combined damage and plasticity formulation was presented that is suitable for modeling of the unloading response of an interface ranging from full damage to full plasticity, while it introduces a minimum number of model parameters that can be experimentally determined. The unloading model can be used with the existing mixed-mode cohesive zone laws that describe the interface loading behavior. The relevance and applicability of the unloading model was demonstrated, in combination with the existing improved Xu-Needlemanmixed mode cohesive law, by modeling the observed combined damage-plasticity unloading response of the above-mentioned glue interface system. In addition, a procedure to identify the model parameters has been presented. Permanent deformation of the sample structure often occurs during delamination tests, particularly, if the layers forming the interface are ductile and the interface is strong. Therefore, accurate determination of the interface fracture toughness requires identification and separation of the contribution of structural plasticity to the total energy dissipation, taking into account the presence of plasticity mechanisms within the fracture process zone at the interface that contribute to the interface fracture toughness. To this end, a semi-analytical approach accounting for the structural plasticity in the sample layers was developed, in order to obtain an accurate value of the interface fracture toughness in a mode I experiment. The approach was numerically verified by employing a finite element model with cohesive zone elements (at the interface). The proposed approach was experimentally assessed by characterizing the interface fracture toughness of industrially relevant copper lead framemolding compound epoxy (CuLF-MCE) structures with different layer thicknesses. In summary, the combined application of in-situ MMMB experiments, the analytical procedure to determine the CERR, and the cohesive zone model with the parameter identification procedure allows for accurate characterization of the delamination mechanism(s) and prediction of the interface mechanics. As a demonstration, industrially relevant coated CuLF-MCE and uncoated CuLF-white molding compound (WMC) interface systems have been characterized in detail using the developed experimental tools