Modeling And Scheduling Of Cluster Tools With Chamber Cleaning Operations

With the increase in wafer diameter,traditional mass production modes can no longer meet the requirements of complex processes such as etching,photolithography,and chemical vapor deposition.Cluster tools adopt single-wafer processing technology,which is a automated device with reconfigurable,high flexibility,high wafer productivity and yield.It has been widely used in wafer production.When cluster tools process wafers,high temperature environment and strong acid and other chemical substances are often accompanied in the process chambers.In order to avoid damage to the surface of the wafer,the fab often cleans the chambers to remove excessive chemical residues.For the chamber cleaning method of cluster tools,regular cycle chamber cleaning and detectionbased chamber cleaning are generally adopted.Besides chamber cleaning constraints,the scheduling of cluster tools is also subject to many constraints,such as wafer residency time,robot type,wafer revisiting,and processing operation status of the cluster tools: start-up process,steady state and close-down process.The above constraints result in a high degree of complexity in the scheduling of cluster tools.In fact,the scheduling problem of cluster tools is an NP-hard problem.Thus,it is of great significance to seek effective scheduling and control methods of cluster tools.The semiconductor manufacturing industry is a capital-intensive industry.Effective scheduling methods can make the production line run smoothly,improve production efficiency and yield,and reduce unnecessary waste.There is an important practical application value for semiconductor companies.For the cluster tools scheduling problem considering the chamber cleaning operations,the following research contents:(1)Considering 1-period of cleaning operations and wafer residency time constraints,the startup process and close-down process scheduling problems of single-arm cluster tools are studied respectively.Firstly,a transient scheduling rule with 1-period of cleaning operations is proposed.Based on this rule,the robot's activity sequence in the transient process is described.Then,Resourceoriented Petri net is used to model and describe the transient of single-arm cluster tools and a rule of transition firing that can avoid deadlock is introduced.Secondly,according to the transient time characteristics of the system and considering different scheduling situations,a corresponding linear programming model a linear programming model for transient scheduling is established.Finally,examples are given to verify the effectiveness of the proposed scheduling model.(2)Using the existing scheduling approach to schedule single-arm cluster tools with detectionbased chamber cleaning operations and wafer residency time constraints,which causes semi-finished wafers accumulation problems Based on this,this paper proposes an improved scheduling approach.Firstly,it analyses the schedulability conditions of the single-arm cluster tools,when the chamber cleaning occurs in the system.Then,according to the characteristics of the wafer flow parameters of the process where the chamber cleaning occurs,two scheduling cases are proposed.Secondly,according to the idea of loading partial wafer to delayed process and the sequence of robot activities under steady-state scheduling,two scheduling rules are proposed and the corresponding scheduling algorithms are derived.If it is schedulable,an optimal scheduling can be obtained through the scheduling algorithms.Finally,illustrative examples are given to show the effectiveness of proposed method.(3)A novel scheduling approach is proposed for dual-arm cluster tools with detection-based chamber cleaning operations and wafer residency time constraints.Firstly,according to the characteristics of swap strategy,the basic activity unit of the dual-arm robot is defined,and the schedulability conditions of the system are given.Then according to the characteristics of wafer flow parameters,two scheduling cases are proposed.Secondly,for different scheduling cases,the scheduling rules that can avoid system deadlock are given,and the algorithm that can obtain the optimal scheduling of the system is deduced.Finally,examples are given to show the application of proposed approach.