Optimizing For Wafer Delay Of Single-armed Cluster Tools With Schedule Algorithm

As the continuous development of semiconductor industry,wafer manufacturing level is rapidly updated and the technical requirements are advanced and complex.Cluster tools are highly integrated and automated equipment which can effectively guarantee the quality of wafer fabrication and improve the production efficiency.With the cleaning process and coating process of wafer,there are have strong corrosive acid and high temperature environment,respectively.In order to avoid to the damage of wafer surface,a wafer should be unloaded from processing module as soon as possible after it completed.Wafer delay is the waiting time in processing module between the completion of the wafer processing until it is unloaded.Time delay has a negative effect on the wafer quality and it is reduced by reasonable control and schedule of robot.Thus,the problem for time delay of single-armed cluster tools is researched in this paper.It can not only optimize the wafer product quality,but also provide theoretical support for the scheduling and control of cluster tools.The content is as follows.(1)In order to analyze the operation of the single-armed cluster tools,the system model should be built first.This paper describes the operation process and task sequence of the single-armed cluster tools under steady state and the Petri net model is built.Meanwhile,the token and transition in the model are timed to make the model suitable for the dynamic process of system.Finally,the control strategy is proposed to solve the potential deadlock situation in the Petri net model.(2)When cluster tools are running under steady state,each system module works in parallel.In order to study the dynamic characteristics of discrete events in the system during operation,this paper analyzes temporal properties of steady scheduling for single-armed cluster tools based on the Petri net model.From the point of view of step,the production cycle of each step is the same.From the point of view of the robot,the calculated expression of wafer residency time and robot cycle time is obtained by analyzing the tasks sequence of the robot.Further,the upper of robot waiting time each step is obtained under steady state.In addition,for the conventional backward strategy and earliest strategy,the distribution rule of robot waiting time under the optimal cycle of the system is proposed and proved.(3)Wafer delay is determined by the scheduling task of the system.In order to control robot reasonably and reduce the wafer delay in the process of system operation,it is necessary to find out the influencing factors of wafer delay.First,the expression of wafer delay is derived.The objective function and constraint conditions are explained by the mathematical programming model and the different influence of the robot waiting time for wafer delay is discussed.Then,A priority rule of robot waiting time allocation is proposed by the influence degree of waiting time.Finally,a scheduling algorithm for single-armed cluster tools is developed based on this priority rule and a robot waiting time sufficiency theorem is proposed.The theorem is used to check whether the waiting time can offset all the wafer delay in system.If the system has enough robot waiting times,total wafer delay of system is zero when the scheduling strategy obtained by the algorithm runs to a steady state.If there are have no enough robot waiting times,the schedule strategy can effectively reduce time delay.(4)In order to check the feasibility of the scheduling algorithm,the algorithm is finished by Matlab software and it solves the steady state scheduling strategy.The operation of cluster tools is simulated under steady state by Gantt chart and illustrative examples are given to its feasibility.As the result shown that under the same conditions,compared with the conventional backward strategy and earliest strategy,the schedule method obtained by the algorithm can not only ensure the system under in the optimal period,but also have shorter time delay.