量測時間延遲和虛擬量測系統對逐批控制之影響分析以及新穎逐批控制技術之推導

Performance analysis of run-to-run controllers subject to metrology delay and virtual metrology and the development novel run-to-run control approach

吳明峰(Ming Feng Wu)

指導教授 ： 鄭西顯

國立清華大學/工學院/化學工程學系/博士(2008年)

https://doi.org/10.6843/NTHU.2008.00148

摘要

本論文主要針對半導體製程生產時所面臨的兩大問題:晶圓量測時間的延遲以及在半導體特有的多量少樣混合產品。根據上述問題，提出修正EWMA控制器使得產品品質可以滿足產品規格。
首先，我們直接以數學模式來分析量測時間的延遲對控制上的影響，提供一個控制上的指導方針，並且使用VM (Virtual Metrology) 來決解延遲之問題；並且在化學氣相沉積製程中加以分析各個建模方法對VM效果之比較，發現以Stepwise最佳之方法。
而在混合產品生產的問題，主要是以統計學之方法ANCOVA來解決，利用ANCOVA可以分析出各個產品對機台的特性，加以修正誤差項，來達到提升控制效能和品質。

關鍵字

逐批控制 ； EWMA 控制器 ； 半導體 ； 化學氣相沉積 ； 混合產品生產 ； 虛擬量測系統 ； 逐步迴歸分析 ； 共變異數分析 ； 時間序例分析

並列摘要

Two problems encountered in many semiconductor manufacturing are metrology delay caused by the equipments and mixture products on producing. For this thesis, we develop some controllers to eliminate the influences of two problems.
First, focusing on the metrology delay problem, we deriver an analytic formula of the performance for a single EWMA controller and provide the guideline to reduce the effects of metrology delay system. Additionally, we provide VM (Virtual Metrology) system to solve it and analyze the performance of VM of modeling methods. Finally, we suggest Stepwise method to plug in VM system.
Mixed-product production can reduce amount of cost and time in semiconductor manufacturing. However, each product has a different I-O model and it is a big challenge to run-to-run control. In order to solve it, we use an ANCOVA model to deal with the characteristic of mixed products in different platforms, and to correct the modeling error due to mixed products in order to impove the controller performance and quality.

並列關鍵字

Run-to-Run ； Metrology delay ； CVD ； CuCmp ； Mixed-product production ； Virtual Metrology ； Stepwise Regression ； ANCOVA ； Time Serise ； Semiconductor manufactory ； EWMA controller

參考文獻

1. Sachs, E., Guo, R. S., Ha , S. and Hu, A. and Ingolfsson A. (1991). “Process control system for VLSI fabrication.” IEEE Transactions on Semiconductor Manufacturing, 4, pp. 134-143.

2. Sachs, E., Hu, A. and Ingolfsson A. (1995). “Run by Run Process Control: Combining SPC and Feedback Control”. IEEE Transactions on Semiconductor Manufacturing, 8, pp. 26-43.

3. Tedder L. L. and Rubloff G. W.( 1996), ”Dynamic rate and thickness metrology during poly-Si rapid thermal chemical vapor deposition from SiH4 using real time in situ mass spectrometry”, J. Vac. Sci. Technol. A 14(2), pp. 267-270.

4. Moyne, J., E.D. Castillo, and A.M. Hurwitz (2001). Run-to-Run Control in Semiconductor Manufacturing. CRC Press, Florida.

5. Castillo E. D.(2002), Statistical Process Adjustment for Quality Control. John-Wiley and Sons, New York.