多變量自適應控制應用於半導體R2R製程

A Multivariate Self-tuning Controller for Run-to-Run Semiconductor Manufacturing Process

任志宏(Chih-Hung Jen)

指導教授 ： 江行全

元智大學/工程學院/工業工程與管理學系/碩士(2001年)

摘要

在最近幾年R2R控制技術的發展常應用於多變的半導體產業，而R2R控制的原理整合了反應曲面、工程製程管制及統計製程管制等基礎理論，其控制的主要目標是透過可控變數的調整使製程反應值儘可能的接近目標值。而本研究的焦點是探討MIMO製程Self-tuning（ST）控制於半導體R2R製程的應用，在控制的過程中除了補償製程中符合ARIMA模型的雜訊項外，對於動態製程中自我迴歸項、製程飄移以及製程偏移老化的現象也一併考量並使製程受控制。
ST控制能夠提供on-line的參數估計及控制，在控制過程中利用遞迴最小平方法來估計on-line製程的參數值。所以多變量ST控制的操作策略是希望偏離目標值及可控變數調整幅度越小使得總成本最小化，並依此概念獲得下一期可控變數應調整多少使得製程能夠達到控制的目標。經由本論文研究的結果顯示，即使製程為非線性模型，論文中所提改善演算法的ST控制比線性模型逼近控制及OAQC系統控制有更好的效果。
最後本論文使用實務半導體製程中化學機械研磨的案例，經由實務製程控制的兩個品質特徵值（研磨率和晶圓內不均勻程度）來評斷本論文所提出的改善方法是可應用於實際製程。

關鍵字

ST控制 ； R2R控制 ； MIMO製程 ； 半導體產業

並列摘要

During recent years, “Run-to-Run” (R2R) control techniques have been developed and used to control various semiconductor manufacturing processes. The R2R control methodology combines response surface modeling, engineering process control (EPC), and statistical process control (SPC). The main objective of such control is to manipulate the recipe so as to maintain the process output of each run as close to the nominal target as possible. The primary focus of this research is on the multiple- input-multiple-output (MIMO) control for self-tuning control of R2R processes. The controller compensates for a variety of disturbances frequently encountered in semiconductor manufacturing, that is, a structured noise of an ARIMA form. The controller also compensates for system dynamics, including autocorrelated responses, deterministic drifts, process shifts, and process gains.
Self-tuning controllers are developed to provide on-line parameter estimation and control. A recursive least squares (RLS) algorithm is normally employed to provide on-line parameter estimation to the controller. So, this control strategy used in a self-tuning controller applies the principle of minimizing total cost, in a sense of an expected off-target and controllable factors adjustment, to obtain a recipe for the next run. It is shown via experimental study even if control model is nonlinear, the self-tuning controller algorithm presented herein can offer better control performance for R2R applications as compared to those of the control action of linear approach of self-tuning controller and the optimizing adaptive quality controller (OAQC) module.
At last, a relevant application to Chemical Mechanical Planarization (CMP) in semiconductor manufacturing, a critical step involving two quality characteristics (removal rate and within-wafer nonuniformity), is used to illustrate the proposed controller.

並列關鍵字

Self-tuning control ； Run-to-Run control ； Multiple input multiple output process ； Semiconductor manufacturing

參考文獻

Alwan, L. and Roberts, H. V. (1988).“Time-Series Modeling for Statistical Process Control,”Journal of Business & Economic Statistics, Vol. 6, No. 1, pp. 87-95.

Baxley, R. V., JR. (1994).“Applications of the EWMA for Algorithmic Statistical Process Control,”Quality Engineering, Vol. 7, No. 2, pp. 397-418.

Boning, D. S. and Mozumder, P. K. (1994).“DOE/Opt: A System for Design of Experiments, Response Surface Modeling and Optimization Using Process and Device Simulation,”IEEE Transactions on Semiconductor Manufacturing, Vol. 7, No. 2, pp. 233-244.

Box, G. E. P. and Jenkins, G. M. (1974). Time Series Analysis, Forecasting, and Control, Revised edition, Holden Day, Oakland, CA.

Box, G. E. P. and Kramer, T. (1992).“Statistical Process Monitoring and Feedback Adjustment - A Discussion,”Technometrics, Vol. 34, No. 3, pp. 251-267.