Thesis (M. S.)--Mechanical Engineering, Georgia Institute of Technology, 2009. / Committee Chair: Sitaraman, Suresh; Committee Member: Peak, Russell; Committee Member: Ume, Charles

The current trend in electronics manufacturing is to decrease the size of electronic components while attempting to increase processing power and performance. This is leading to increased interest in thinner printed wiring boards and finer line widths and wire pitches. However, mismatches in the thermomechanical properties of materials used can lead to warpage, hindering these goals. Warpage can be problematic as it leads to misalignments during package assembly, reduced tolerances, and a variety of operational failures. Current warpage prediction techniques utilize isotropic volume averaging to estimate effective material properties in layers of copper mixed with interlayer dielectric material. However, these estimates do not provide material properties with sufficient accuracy to predict warpage, as they contain no information about the orientation of the copper traces. This thesis describes the development of a new technique to predict the warpage of a particular substrate. The technique accounts for both the trace pattern planar density and planar orientation in determining effective orthotropic material properties for each layer of a multi-layer substrate. Starting with the trace pattern image, this technique first divides the trace pattern into several smaller areas for a given layer of the substrate and then uses image processing techniques to determine the copper percentage and average trace orientation in each small area. The copper percentage and average trace direction orientation are used in conjunction with the material properties of copper and the dielectric material to calculate the effective orthotropic material properties of each smaller area of the substrate. A finite-element model is then created where each layer is represented as a concatenation of several small areas with independent directional properties, and such a model is then subjected to sequential thermal excursion as seen in the actual fabrication process. The results from the models have been compared against experimental data with a great degree of accuracy. The modeling technique and the results obtained clearly demonstrate the need for the proposed subdivisional orthotropic material property calculations, as opposed to homogeneous isotropic properties typically used for each layer in computational simulations, as these more accurate directional properties are capable of predicting warpage with higher accuracy.