整合線上實驗與控制理論於製程模型之變動

Integrate on-line experiment and the control methods for changes in a dynamic model

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/博士(2007年)

https://doi.org/10.6838/YZU.2007.00001

摘要

在過去的十年間，R2R控制已被廣泛應用在不同的半導體製程。然而，如此控制方式有一個缺點，那就是當製程突然產生較大的改變時，由於製程模型偏離而無法滿足控制的需求。基於這個原因，本論文將反應曲面、線上實驗以及R2R控制的想法整合成一個新的製程控制系統。研究的主要目標是改善動態模型的參數預測，使得控制系統能夠更有效的計算可控變數調整。在系統的操作上，是透過遞迴最小平方(RLS)估計以及合併(joint)參數界限值去評估製程是否產生一個較大的改變。假如評估值超出參數界限值，則可控變數調整將移往下一期控制的最佳點，並在最佳點附近透過逐漸放大實驗區間的方法來收集資料以及使用最小平方估計(LSE)更新製程模型。接下來，控制系統使用成本函數(the cost function)和BFGS演算法計算下一期可控變數調整的最佳解。從模擬結果發現，這個新的控制系統比Self-tuning控制擁有較佳的控制結果。此外當製程屬於較輕微的動態改變時，這個系統也比蛻變作業(EVOP)更容易操作。最後在實務的應用上，本論文透過兩個半導體關鍵製程的模擬(機械化學研磨(CMP)和閘極蝕刻(polysilicon gate etching))去說明本研究方法的可行性。

關鍵字

反應曲面；線上實驗；蛻變作業(EVOP)；R2R控制；Self-tuning控制；遞迴最小平方估計(RLS)；BFGS演算法

並列摘要

During the past decade, a variety of ‘run-to-run’ (R2R) control schemes have been proposed and investigated extensively under various semiconductor manufacturing methods. However, such control has a problem when it is suddenly faced with a larger process change that does not satisfy the control requirement. In view of this, a new process control framework is proposed that integrates response surface modeling, on-line experiment, and R2R control ideas. The primary objective of this study is to improve dynamic model parameter prediction, enabling more effective optimized recipe calculation. The recursive least squares (RLS) algorithm is used to evaluate changes in process parameters to establish the degree of control for the next period. If the evaluated parameter values exceed a joint parameter threshold, the recipe moves to a new optimum point. This moment, which will apply the design of experiment concept continuously to collect process data in the experimental range, uses this data with the least squares error (LSE) method to estimate the new model parameters. The proposed control strategy uses the minimized total cost principle (the cost function form includes an expected off-target and controllable factors adjustment) and applies the Broyden-Fletch-Goldfarb-Shanno (BFGS) algorithm to obtain a recipe for the next period. Simulation studies show that the proposed system has better control performance than the traditional self-tuning controller. When the model is provided with a slight variation, operating and monitoring are easier than using evolutionary operation (EVOP).
In the relevant chemical mechanical planarization (CMP) and polysilicon gate etching applications in semiconductor manufacturing, two critical chip fabrication steps are used to illustrate the proposed control system in a dynamic process.

並列關鍵字

Response surface modeling ； On-line experiment ； Evolutionary operation (EVOP) ； Run-to-run (R2R) control ； Self-tuning control ； Recursive least squares (RLS) algorithm ； Broyden-Fletch-Goldfarb-Shanno (BFGS) algorithm

參考文獻

Alwan, L. and Roberts, H. V. (1988) Time-Series Modeling for Statistical Process Control. Journal of Business & Economic Statistics, 6, 87-95.

Asprey, S. P. and Macchietto, S. (2000) Statistical Tools for Optimal Dynamic Model Building. Computers and Chemical Engineering, 24, 1261-1267.

Asprey, S. P. and Macchietto, S. (2002) Designing Robust Optimal Dynamic Experiments. Journal of Process Control, 12, 545-556.

Åström, K. J. (1970) Introduction to Stochastic Control Theory. (New York, NY : Academic Press).

Åström, K. J., Borisson, U., Ljung, L., and Wittenmark, B. (1977) Theory and Applications of Self-Tuning Regulators. Automatica, 13, 457-476.