Dynamic deployment modeling tool for photolithography WIP management

In semiconductor manufacturing, according to Marcoux et al. (1999) tool deployment has been identified as a key factor driving capacity loss and lower operational efficiency. In most cases, the losses are uncovered by analysis of Cycle Time data and investigation of specific tool performance. For the photolithography sector, this feedback approach often highlights problems after they may have already past or have been fixed. This paper will discuss a feed forward model for managing deployment of a large fleet of photolithography tools. This model predicts tool loading using existing tool planning parameters, actual and forecast wafer start data and extensive turn-around-time matrices. The model provides a portable tool with immediate readout of various loading scenarios. The deployment decision process makes use of these simulations. The model output comes in the form of graphs and tables that can summarize load by tool, tool groups, resist groups, technologies, and levels at various time slices. The output identifies where tool qualifications or additional resists may be needed, and deployment adjustments for WIP balance is warranted. These changes prevent operational efficiency loss and maintain cycle time performance.