Cleaning Plan Optimization for Dual-Armed Cluster Tools With General Chamber Cleaning Periods

A cluster tool, widely used for semiconductor wafer fabrication, consists of several single-wafer processing chambers and a wafer handling robot. A chamber for critical processes is periodically cleaned for removing chemical impurities to reduce quality risk due to extreme circuit width shrinkage. We examine the problem of determining a cleaning plan for dual-armed cluster tools with general chamber cleaning periods $k_{i} > 1$ for chamber $i$ using the popular swap sequence to minimize the cycle time. We derive conditions for which the popular swap sequence minimizes the tool cycle time regardless of the cleaning plan. For the other cases, we develop a cleaning rule named DGC(Dispersing and Gathering Cleaning) that disperses cleaning operations along the robot cycle and gathers the cleaning operations that are interlaced along the robot cycle. For parallel flows with a single process step, we develop a closed-formula for the cycle time and prove that DGC minimizes the cycle time for the swap sequence. We also present conditions under which the swap sequence with the cleaning rule achieves the minimum cycle time compared to all other sequences and cleaning plans. For serial wafer flows, we show that DGC minimizes the cycle time when the cleaning period and the cleaning time are the same for all process steps. We also show by experiments that the proposed DGC effectively reduces the cycle time for the other general cases. Note to Practitioners—Our proposed cleaning rule can significantly improve the tool cycle time and regulate wafer flow times. It is easily applied or adapted to and effective for most wafer flow patterns, sequences, and tool architectures.