A schedule of cleaning processes for a single-armed cluster tool

A conventional robot sequence of a single-armed cluster tool pursues a backward strategy for efficiency; the strategy is optimal when identical work cycles are repeated. However, we cannot guarantee the strategy is optimal under some abnormal situations, such as breakdowns or during auto-cleaning (periodic cleaning for preventive maintenance). In this paper, we focus on undesirable but predictable auto-cleaning, which cleans the chambers of a cluster tool each time a predefined number of wafers has been processed to prevent deterioration of wafer quality caused by residual chemicals. For auto-cleaning, we suggest alternative cleaning frequencies of a single-armed cluster tool that periodically clean the inside of process modules based on the Theory of Constraints. Additionally, we derive its heuristic robot scheduling from Petri-net models, because finding the optimal robot sequence in real time is impossible. Finally, we examine the effectiveness of our suggestions by simulation.