Improvement of mold ability and wire sweep in semiconductor devices using DOE statistical approach

Molding, also known as encapsulation in semiconductor industry, is one of the main processes in semiconductor manufacturing. In semiconductor industry, encapsulation is usually done using transfer molding due to very high accuracy of transfer molding tooling and low cycle time of the molding process. In order to optimize and obtain the perfect molding process, molding simulation is necessary to check the mold flow and possibilities of voids and air trap inside the mold tools. However, simulation alone could not guarantee the perfect molding process. Usually simulation have some limitation such as to match the exact mold compound properties, and when this happen few potential defects may be produced such as incomplete mold, crack mold, mold chipped, uneven surface and many more. These defects are also called mold ability issues. Normally, the mold ability defect will result on drop of function performance or reliability issue during application of the product. Another defect that can be produced if the parameter is not optimized is wire sweep. Wire sweep is defined as the ratio of the maximum wire deviation or deformation to the wire span. In other words, if observed the wire from top view, the straighter the wire, the less wire sweep index ratio obtained. This research aims to investigate and optimize molding parameter by using Design of Experiments approach to achieve zero mold ability issue and good wire sweep index, which will improve the molding process yield and further reduce the molding defect produced. Effects of molding parameters to the mold ability and wire sweep performance will be study and based on the analysis, best molding parameter will be defined and recommended to minimize the yield loss and improve the company profitability. Generally, every molding parameter (Mold Temperature, Transfer Pressure, Transfer Time) having effects and interaction to mold ability and wire sweep output. The statistical calculation was done using Cornerstone software, and based on the analysis, the best parameter was suggested. This study only focused on mold parameter optimization, so in order to further increase the robustness, it is recommended to study other related factors such as mold compound characteristics and mold tool design