應用小波理論於薄膜製程之射頻功率故障偵測

RF Power Fault Detection of Thin Film Process Using Wavelet Transform

陳源杰(Yuan-Chie Chen)

指導教授 ： 張耀仁

中原大學/工學院/機械工程研究所/碩士(2005年)

https://doi.org/10.6840/cycu200500351

摘要

電漿增強化學氣相沉積在薄膜電晶體液晶顯示器的陣列製程中主要應用於非晶矽與氮化矽的薄膜沉積，需要搭配射頻系統來產生電漿源。穩定的射頻功率是電漿源生成的基本條件，然而在薄膜沉積的製程中，機台經常發生的故障情形就是射頻功率的損失，導致成膜不全和影響下一批次的製程，降低生產良率與設備產能。
小波基底函數具有自動縮張的性質，可時-頻區域化分析的優點，以及對信號自適應的特性，因此本研究針對製程的相關參數進行信號處理，以即時移動視窗的觀念將參數作線上離散小波轉換，並且配合適應性臨界值來增進殘值估算的強健性。探討的機台參數包括氣體流量、反應室壓力和射頻功率等，其中反射功率對射頻功率損失的議題可以達到較為有效的故障偵測目標。

關鍵字

射頻功率損失 ； 小波轉換 ； 故障偵測

並列摘要

Plasma enhanced chemical vapor deposition (PECVD) is an important process in the array engineering of TFT-LCD manufacturing for thin-film deposition of amorphous silicon or silicon nitride. The plasma is generated by a RF system. Therefore, stable RF power is the basic operation requirement for high quality thin-film deposition. Unfortunately, the RF power loss is a very common fault during the deposition process and always leads to defect films. Moreover, it may affect the next process run and reduce the yield and throughput.
Some advantages of the wavelet basis functions include the variable aspect ratio, the time-frequency localized analysis, and the self-adaptive characteristics to signals. In this study, the process parameters were treated with on-line discrete wavelet transform using a moving window. Furthermore, an adaptive threshold was calculated for the robust residual evaluation. Several process parameters, including the gas flow rate, reactor pressure, and RF power, were collected and investigated. Consequently, among the process parameters, the reflection power was found to be the effective factor for detecting the fault of RF power loss.

並列關鍵字

fault detection ； RF power loss ； wavelet transform

參考文獻

[2] M. Quirk and J. Serda, “Semiconductor Manufacturing Technology,” Prentice Hall, 2001.

[5] B. Kim and G. S. May, “Real-time diagnosis of semiconductor manufacturing equipment using a hybrid neural network expert system,” IEEE Trans. Comp. Pack. Manufact. Technol., v 20, n 1, p 39-47, Jan, 1997.

[6] B. Kim and C. Lee. “Monitoring plasma impedance match characteristics in a multipole inductively coupled plasma process control,” J Vac. Sci. Technol. A, v 18, n 1, p 58-62, Jan, 2000.

[7] B. Kim, “Radio frequency transient match monitoring using a fast Fourier Transform for plasma diagnosis,” J. Vac. Sci. Technol. A, v 19, n 2, p 700-702, March/April, 2001.

[8] B. Kim and W. Choi, “Discrete wavelet monitoring of plasma impedance matching for process control,” IEEE International Symposium on industrial Electronics, v 1, p 171-175, 2001.