Here is the abstract you requested from the imaps_2019 technical program page. This is the original abstract submitted by the author. Any changes to the technical content of the final manuscript published by IMAPS or the presentation that is given during the event is done by the author, not IMAPS.
|Mechanical Modeling and Continuous Process Improvement|
|Keywords: Finite Element Modeling, Irregular Rough Structures, Crack Prediction|
|Short design cycles for integrated circuits and packages drive the need for efficient problem solving and rapid results. Improved mechanical modeling software and increased computing power have taken these computation-heavy tools and made them versatile enough to support main- stream, real-time production needs. The utility of these tools has been significantly improved by simplified work flows to create detailed geometries and complex assemblies, improved mesh generation algorithms, and solve time reduction. Mechanical modeling software has a wide range of application which traditionally has been focused on design of large structures. Despite their general applicability, these tools have not been optimized for microelectronics in terms of absolute dimensions, fine structure count, and range of scale from the smallest to the largest component. Finding solutions to these problems pays off in fewer design cycles and significant process yield improvements. This paper will show multiple examples of process-induced stress, driven by material properties and manufacturing. They have been created using a variety of FEM tools, including Ansys and Abaqus. CASE #1 - Rough Thin Film geometry challenges Traditional mechanical modeling deals primarily with idealized geometries with smooth surfaces and simple lines or curves. Simulation of less-than- ideal structures and surfaces has been challenging in terms of geometry creation and computational limitations. Nevertheless, there are increasing numbers of cases in which random surface features are created on the same scale as the structure geometry. In such cases mechanical modeling based on idealized geometry may fail to predict component performance. Due to the growing popularity of 3- D printing, many 3-D drafting and FEM tools have added support for import and manipulation of faceted geometry files, e.g. STL format. Various commercial and freeware tools can be used for generating the rough surface geometries. These faceted geometries can then be cleaned and imported into the FEM solving tools where they are meshed and solved like standard geometries, albeit with higher element count. One such example considered is the deposition of thin brittle films over thick electroplated metal. The brittle film builds on the underlying metal roughness and magnifies the size of the random surface features. Since the brittle film coefficient of thermal expansion (CTE) is much lower than that of the underlying metal, large stresses develop that can cause the film to crack. To improve manufacturability, variations in metal layout and film thickness are considered. Results from the simulation experiments will be presented in which the stress functional dependence on layout and film thickness is shown. Benefits of using explicitly modeled rough surfaces are discussed, including comparison to standard techniques and improved acceptance of simulations that look like the actual geometry. CASE #2 - Device fracture prediction on epoxy film structures due to intrinsic stresses, processing and testing. Photosensitive polymers are found in many microelectronic lithography applications as well as components in final assemblies. For these applications intrinsic film stresses arise during fabrication. Subsequent process steps may use materials with different CTE. The thermal mismatch resulting from temperature changes plus intrinsic stresses generates thermo-mechanical stresses that can lead to structural failure due to delamination and cracks. Since electronic components must endure extremely challenging assembly and test conditions, predicting the performance of these materials is essential to producing robust products. The extended finite element method (XFEM) is an extension of the classical finite element method (FEM) approach that allows modeling the initiation and propagation of various discontinuities (including delamination and cracks) without the need to re-mesh the critical region. This paper will show an example in which XFEM is applied to a capture crack initiation and propagation in epoxy film structures due to the intrinsic stresses of fabrication and assembly, and the extremes of thermal cycling. The results were able to replicate an observed failure mode with a high degree of accuracy.|
|Mercedes T. Hernandez, Sr. Development Engineer