Scheduling Control Of Mixed-processing With Multi-variety Wafers In Dual-armed Cluster Tools

Semiconductor manufacturing is basic industry to support the information industry.The wafer manufacturing process is complex,and traditional wafer fabrication has been difficult to meet complex processing requirements.The cluster tools provide a closed,vacuum,efficient and reconfigurable environment for single wafer production to ensure strict processing requirements,which is conducive to improve the quality of semiconductor products.In addition,with the increase of wafer size and requirements for customers,wafer fabrication tends to multi-variety and small batch processing.The cluster tools has been widely applied to the manufacture of large size wafers,which is beneficial to meet the productivity demand for customers.With the increasing varieties of wafers,different types of wafers are frequently switched during the scheduling in cluster tools,which results in more difficulties in system scheduling,longer switching periods and increasing in manufacturing costs.Therefore,to effectively control the operation of cluster tools is of great industrial application value.In order to improve the operation efficiency of cluster tools and reduce the adapting period for cluster tools during wafer switching,the control problem of dual-armed cluster tools is concerned,the content is as follows.(1)First,we introduce the scheduling processes of dual-armed cluster tools for multi-variety wafer fabrication,and the definition and characteristics of Petri nets.Based on the swap strategy,a Petri net model of dual-armed cluster tools is established in steady state.According to the established Petri net model,we further analyze the activity sequence and propose a deadlock free control strategy.After analyzing the wafer processing time and residency time constraints,the necessary and sufficient conditions of system activity are obtained.(2)According to the characteristics of system activity cycle and transitions firing sequence in the Petri net model,furthermore,the wafer processing cycle and robot activity cycle can be obtained in the steady state.Aiming to the situations of sharing and without sharing processing modules,we analyze the relationship between the system activity cycle,wafer processing cycle and robot activity cycle to determine the schedulability of different time distribution,finally the theorems are proved correct.(3)In order to directly describe the scheduling process in cluster tools from start-up transient to steady state,we take sharing and without sharing processing modules as instance.Two eM-Plant models of dual-armed cluster tools are established for multi-variety wafer processing.During the operation of simulation models,the entry time of wafers in each module is recorded,then the schedulability of the system is obtained by analyzing the residence time constraints of wafer and time.The research work shows that the obtained schedulability of multi-variety wafer mixed-processing can be well used for dual-armed cluster tools,as well as the established eM-Plant simulation models,which has a certain application value.