Image-based interface characterization with a restricted microscopic field of view

Abstract

Accurate characterization of adhesion properties in microelectronic systems is challenging due to (1) the far-field load application that often falls outside the microscopic field of view, (2) the ultra-small loads associated with specimen deformation, and (3) the load-case and specimen dependent interface response. To overcome these challenges, a generic method based on Integrated Digital Image Correlation (IDIC) is proposed, which identifies cohesive zone model parameters (of an arbitrary model not intrinsic to the identification method), by correlating images of a delamination process from a restricted field of view at the microscopic scale, whereby far-field loading data cannot be exploited.To quantify the effects of potential error sources on the performance of the proposed IDIC-routine, virtual experimentation is first conducted. Inaccurate application of boundary conditions in the FE-model of IDIC is thereby shown to be the most critical source of error. Subsequently, a real double cantilever beam (DCB) experiment has been analyzed as a well-defined test-case for characterization of adhesion properties. Since the Young’s modulus of the bulk material is generally well known, the imaged, elastically deforming bulk material acts as a force sensor. External load measurement can therefore be omitted from the identification process, thereby rendering the interface identification method independent of the particular test method. The implemented IDIC-algorithm is shown to be robust for accurately identifying the two cohesive zone parameters of interest: the work of separation Gc and the critical opening displacement δc.