半導體製程隨著摩爾定律推進至奈米世代，蝕刻劑作為半導體製程的關鍵材料，對於純度要求日益嚴苛。本研究個案公司在過往生產實務中，特定雜質（雜質A）之濃度不穩定。傳統依賴工程師經驗法則與試誤法的調機模式，不僅耗費大量時間成本，更因無法釐清多重參數間的複雜交互作用，導致製程能力不足。
為解決上述問題，本研究導入實驗設計法（Design of Experiments, DOE）進行系統性的參數最佳化。首先透過數據分析與關鍵參數篩選，鎖定製程溫度與製程流量為關鍵控制因子。研究採取循序漸進策略，第一階段執行全因子實驗並加入中心點，檢定結果顯示曲率顯著（P < 0.05），證實製程變數間存在非線性關係。遂於第二階段導入反應曲面法（Response Surface Methodology, RSM），採用中央合成設計建構二階多項式預測模型。
研究結果發現，製程溫度與製程流量之間存在顯著的交互作用。雖然提高溫度有助於分離，但易引發霧沫夾帶效應導致雜質上升，此時需搭配較高的流量以發揮迴流洗滌機制。
經由模型數值最佳化求解，確立最佳操作參數。經確認實驗驗證，成品雜質A之平均濃度成功降低，且變異數顯著降低，落於模型預測之95%信賴區間內。本研究證實透過科學化的實驗設計，能有效解析複雜的化工分離機制，取代傳統經驗法則，大幅縮短調機時間並提升產品品質穩定性。

The semiconductor manufacturing industry is advancing fast, shifting purity requirements for wet chemicals like etchants from parts-per-million to parts-per-trillion levels. This research investigates a process optimization case within a leading electronic-grade chemical supplier facing persistent instability in specific metallic "Impurity A" levels. Historical data indicated unpredictable fluctuations often exceeding control limits. Traditional empirical methods and trial-and-error adjustments proved inefficient, failing to elucidate the complex, non-linear interactions governing separation efficiency.
To address these critical quality issues, this study implements a systematic robust parameter design framework utilizing Design of Experiments (DOE). The research employed a sequential strategy, analyzing historical data to identify three critical parameters: two process temperatures and one system flow rate.
A Full Factorial Design first screened for main effects and curvature. Statistical analysis revealed significant curvature (P < 0.05), confirming the relationship between parameters and impurity concentration is inherently non-linear.Consequently, the study advanced to Response Surface Methodology (RSM) using a Central Composite Design (CCD) to map the optimal operating window. The analysis unveiled a significant synergistic interaction between process temperatures and flow rate. Findings suggest that while higher temperatures provide necessary separation energy, they increase the risk of impurity entrainment.
However, optimizing the system flow rate mitigates this by enhancing the internal washing mechanism. Numerical optimization identified specific parameter settings, and confirmation experiments validated that these conditions successfully minimized Impurity A and reduced process variability. This research demonstrates that statistical optimization effectively replaces legacy empirical methods, establishing a stable, high-purity manufacturing process suitable for next-generation applications.

中文文獻
余豐榮. (2013). 田口方法於鋁合金晶粒細化最佳壓縮參數之研究. 科學與工程技術期刊.頁77-85.
李世平. (2010). 電漿化學氣相沉積技術成長超低應力之紫外光可透性氮化矽薄膜. 國立交通大學半導體材料與製程設備學程.
林煜斌. (2006). 鎳鈀金PPF釘架銲線製程的研究. 國立成功大學工程科學系碩士在職專班.
楊沛子. (2018). 反應曲面法設計界面活性劑萃取茶樹精油最佳化之研究. 國立成功大學化學工程學系.
陳寵文. (2007). 以田口方法探討光電廠去光阻劑回收系統運轉之最佳化. 國立交通大學工學院專班永續環境科技學程.
陳彥旻. (2002). 半導體業化學機械研磨廢水回收處理再利用技術研究. 國立成功大學環境工程學系.
英文文獻
Anderson, M. J. (2024). Mixture Design of Experiments (DOE) for Optimal Plasma Etch. Journal of Plastic Film & Sheeting, 41(1), 1–7.
Aslan, N., & Cebeci, Y. (2007). Application of Box–Behnken design and response surface methodology for modeling of some Turkish coals. Fuel, 86(1–2), 90–97.
Bezerra, M. A., Santelli, R. E., Oliveira, E. P., Villar, L. S., & Escaleira, L. A. (2008). Response surface methodology (RSM) as a tool for optimization in analytical chemistry. Talanta, 76(5), 965–977.
Bradley, N. (2007). The Response Surface Methodology. Department of Mathematical Sciences ,Indiana University of South Bend.
Durakovic, B. (2017). Design of experiments application, concepts, examples: State of the art. Periodicals of Engineering and Natural Sciences (PEN), 5(3), 421–439.
Ghanbarzadeh, S., Valizadeh, H., & Zakeri-Milani, P. (2013). Application of Response Surface Methodology in Development of Sirolimus Liposomes Prepared by Thin Film Hydration Technique . BioImpacts, 3(2), 75–81.
Mori, T., & Tsai, S.-C. (2011). Taguchi Methods: Benefits, Impacts, Mathematics, Statistics and Applications. ASME Press.
Nerurkar, T. R. (2015). Design of Experiments on a Semiconductor Plasma Ashing Process: Methods and Analysis. Department of Mechanical Engineering,The University of Michigan-Ann Arbor.
Oehrlein, G. S., Brandstadter, S. M., Bruce, R. L., Chang, J. P., DeMott, J. C., Donnelly, V. M., Dussart, R., Fischer, A., Gottscho, R. A., Hamaguchi, S., Honda, M., Hori, M., Ishikawa, K., Jaloviar, S. G., Kanarik, K. J., Karahashi, K., Ko, A., Kothari, H., Kuboi, N., … Ventzek, P. L. G. (2024). Future of plasma etching for microelectronics: Challenges and opportunities. Journal of Vacuum Science & Technology B, 42(4), 2–44.
Pakpum, C., & Pussadee, N. (2015). Design of Experiments for (100) Si Vertical Wall Wet Etching Using Sonicated NaOH Solution. Applied Mechanics and Materials, 804, 12–15.
Ranga, S., Jaimini, M., Sharma, S. K., Chauhan, B. S., & Kumar, A. (2014). A Review on Design OF Experiments (DOE). International Journal of Pharmaceutical, Biological and Chemical Sciences, 3(1), 216–224.
Sadri Moghaddam, S., Alavi Moghaddam, M. R., & Arami, M. (2010). Coagulation/flocculation process for dye removal using sludge from water treatment plant: Optimization through response surface methodology. Journal of Hazardous Materials, 175(1–3), 651–657.
Shyam Kumar Karna. (2012). An Overview on Taguchi Method. International Journal of Engineering and Mathematical Sciences, 1, 11–18.
Ting, H. T., Abou-El-Hossein, K. A., & Chua, H. B. (2009). Prediction of etching rate of alumino-silicate glass by RSM and ANN. Journal of Scientific & Industrial Research, 68, 920–924.