For the semiconductor packaging industry, the issue of wire sag has not been touched upon because only one single layer of wire bond is constructed along the perimeter of the chip system. Over-sag deflections may not cause any problems. However, driven by the demands for smaller and faster advanced microelectronic devices, modern 3-D and multichip modules have become the solution of choice in delivering higher integration and more multifunctional products to meet consumer needs. For 3-D and multichip packaging, excessive wire sag can lead to wires touching against the lower layer, causing a short-circuit and failure of chip functionality. Therefore, gaining knowledge of wire bond sag theory becomes extremely important for advanced packaging. We propose an empirical model to predict the sag stiffness of a wire bond. The sag stiffness of a wire bond is defined to represent its resistance to the sag drag force occurring during the encapsulation process by the molding flow. A higher sag stiffness of a wire bond gives the premise of less sag deflection, thereby avoiding short-circuiting and/or double wire crosstalk due to the closeness of the top-layer to the bottom-layer wire bonds. The theory of sag regarding semiconductor wire bonding has not been documented, to our knowledge, in the literature. Therefore, to deal with this issue, this paper will primarily focus on the theoretical and numerical analysis of sag deflection of a wire bond. A set of sag experiments is conducted to verify the validity of the sag model hypothesis outlined in the earlier paper. Subsequently, an empirical sag model could be put forward to predict the sag stiffness of a wire bond if the corresponding bond spans, bond heights, and material properties are known. The engineered design of a wire bond can be improved through the use of this sag equation. By making use of the proposed sag model, one can attain not only a first-order estimation of the sag deflection but also omit the tedious procedures of numerical analysis.
Read full abstract