Mixed-mode cohesive zone parameters from integrated digital image correlation on micrographs only

Abstract

Abstract Mixed-mode loading conditions strongly affect the failure mechanisms of interfaces between different material layers as typically encountered in microelectronic systems, exhibiting complex material stacking and 3D microstructures. The integrated digital image correlation (IDIC) method is here extended to enable identification of mixed-mode cohesive zone model parameters under arbitrary levels of mode-mixity. Micrographs of a mechanical experiment with a restricted field of view and without any visual data of the applied far-field boundary conditions are correlated to extract the cohesive zone model parameters used in a corresponding finite element simulation. Reliable or accurate force measurement data is thereby not available, which constitutes a complicating factor. For proof-of-concept, a model system comprising a bilayer double cantilever beam specimen loaded under mixed-mode bending conditions is explored. Virtual experiments are conducted to assess the sensitivities of the technique with respect to mixed-mode loading conditions at the interface. The virtual experiments reveal the necessity of (1) optimizing the applied local boundary conditions in the finite element model and (2) optimizing the region of interest by analyzing the model’s kinematic sensitivity relative to the cohesive zone parameters. From a single test-case, exhibiting a range of mode-mixity values, the mixed-mode cohesive zone model parameters are accurately identified with errors below\xa01%. The IDIC-procedure is shown to be robust against large variations in the initial guess values for the parameters. Real mixed-mode bending experiments are conducted on bilayer specimens comprising two spring steel beams and an epoxy adhesive interface, under different levels of mode-mixity. The mixed-mode cohesive zone model parameters are identified, demonstrating that IDIC is a powerful technique for characterizing interface properties of interfaces, imaged with a limited field of view, as is typically the case in microelectronic applications.