Early identification of inter-process interactions to expedite semiconductor technology development

The period of maximum profitability for semiconductor products is short and occurs early in production, thereby forcing the industry to continually strive to reduce the development time for new technologies. Interactions between the many process steps in semiconductor process flows complicate and significantly lengthen the flows' development by introducing design cycles to adjust for the impact of the interactions. This dissertation focuses on gaining interaction knowledge about arbitrary flows early in their design and scheduling their development so as to minimize the engineering effort involved and total development cycle time.; A qualitative simulation of the process flow is employed to predict the conditions of the wafer state throughout the flow, and an inference engine identifies modeled inter-process interactions that may occur in the flow based upon these wafer conditions. A second simulator then refers to the identified interactions to predict the probable progression of the flow's development for a given schedule of the individual process steps' development activities. With the described system, flow integration engineers can view the interdependencies present in the entire flow and can optimize the flow's development taking these dependencies into account.; A portion of a 0.35 micron technology flow from industry was evaluated as a proof of principle for the system. With 60 interactions codified, PIER identified 192 significant inter-process interactions between the 30 modeled process steps of this flow. During the flow's actual development, too much time was spent redeveloping process steps after they were unsuccessfully integrated with their neighboring steps in the flow. By improving the development schedule, expected reductions in engineering effort on the order of 15% were demonstrated. This percentage increases if otherwise unforeseen interactions are identified or if selected process steps are expedited in accordance with the appropriate sensitivity analyses.; An early understanding of a flow's inter-process interactions will improve the development of the technology in four ways: (a) potential flow design errors are more easily identified, (b) more complete information is provided to the unit process engineers, (c) ramifications of subsequent process changes are identified, and (d) development schedules can be optimized to minimize design cycles.